Free LLM Models Directory
Discover free and open-source large language models for text generation, code completion, and embeddings.
The Ultimate Free LLM Models Directory
Welcome to the most comprehensive directory of free and open-source LLM models. Whether you're a developer, researcher, or AI enthusiast, our curated collection helps you discover powerful language models that you can run locally, fine-tune, or integrate into your applications.
Our directory features models from leading providers including Meta's Llama 2, Mistral AI, Microsoft's Phi, and hundreds more on Hugging Face. All models are verified to have permissive licenses allowing commercial use.
From text generation and code completion to embeddings and multimodal models - find everything you need to build the next generation of AI applications!
Showing 145-192 of 576 models
OpenAI: GPT-5.1-Codex-Mini
LLM Models
GPT-5.1-Codex-Mini is a smaller and faster version of GPT-5.1-Codex. OpenAI: GPT-5.1-Codex-Mini is listed as a large language or generative AI model associated with openai, with the model identifier openai/gpt-5.1-codex-mini. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
MoonshotAI: Kimi K2 Thinking
LLM Models
Kimi K2 Thinking is Moonshot AI’s most advanced open reasoning model to date, extending the K2 series into agentic, long-horizon reasoning. Built on the trillion-parameter Mixture-of-Experts (MoE) architecture introduced in. MoonshotAI: Kimi K2 Thinking is listed as a large language or generative AI model associated with moonshotai, with the model identifier moonshotai/kimi-k2-thinking. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Amazon: Nova Premier 1.0
LLM Models
Amazon Nova Premier is the most capable of Amazon’s multimodal models for complex reasoning tasks and for use as the best teacher for distilling custom models. Amazon: Nova Premier 1.0 is listed as a large language or generative AI model associated with amazon, with the model identifier amazon/nova-premier-v1. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Perplexity: Sonar Pro Search
LLM Models
Exclusively available on the OpenRouter API, Sonar Pro's new Pro Search mode is Perplexity's most advanced agentic search system. It is designed for deeper reasoning and analysis. Pricing is based. Perplexity: Sonar Pro Search is listed as a large language or generative AI model associated with perplexity, with the model identifier perplexity/sonar-pro-search. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Mistral: Voxtral Small 24B 2507
LLM Models
Voxtral Small is an enhancement of Mistral Small 3, incorporating state-of-the-art audio input capabilities while retaining best-in-class text performance. It excels at speech transcription, translation and audio understanding. Input audio. Mistral: Voxtral Small 24B 2507 is listed as a large language or generative AI model associated with mistralai, with the model identifier mistralai/voxtral-small-24b-2507. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
OpenAI: gpt-oss-safeguard-20b
LLM Models
gpt-oss-safeguard-20b is a safety reasoning model from OpenAI built upon gpt-oss-20b. This open-weight, 21B-parameter Mixture-of-Experts (MoE) model offers lower latency for safety tasks like content classification, LLM filtering, and trust. OpenAI: gpt-oss-safeguard-20b is listed as a large language or generative AI model associated with openai, with the model identifier openai/gpt-oss-safeguard-20b. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
NVIDIA: Nemotron Nano 12B 2 VL (free)
LLM Models
NVIDIA Nemotron Nano 2 VL is a 12-billion-parameter open multimodal reasoning model designed for video understanding and document intelligence. It introduces a hybrid Transformer-Mamba architecture, combining transformer-level accuracy with Mamba’s. NVIDIA: Nemotron Nano 12B 2 VL (free) is listed as a large language or generative AI model associated with nvidia, with the model identifier nvidia/nemotron-nano-12b-v2-vl:free. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
MiniMax: MiniMax M2
LLM Models
MiniMax-M2 is a compact, high-efficiency large language model optimized for end-to-end coding and agentic workflows. With 10 billion activated parameters (230 billion total), it delivers near-frontier intelligence across general reasoning,. MiniMax: MiniMax M2 is listed as a large language or generative AI model associated with minimax, with the model identifier minimax/minimax-m2. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Qwen: Qwen3 VL 32B Instruct
LLM Models
Qwen3-VL-32B-Instruct is a large-scale multimodal vision-language model designed for high-precision understanding and reasoning across text, images, and video. With 32 billion parameters, it combines deep visual perception with advanced text. Qwen: Qwen3 VL 32B Instruct is listed as a large language or generative AI model associated with qwen, with the model identifier qwen/qwen3-vl-32b-instruct. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
IBM: Granite 4.0 Micro
LLM Models
Granite-4.0-H-Micro is a 3B parameter from the Granite 4 family of models. These models are the latest in a series of models released by IBM. They are fine-tuned for long. IBM: Granite 4.0 Micro is listed as a large language or generative AI model associated with ibm-granite, with the model identifier ibm-granite/granite-4.0-h-micro. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Microsoft: Phi 4 Mini Instruct
LLM Models
Phi-4-mini-instruct is a lightweight open model built upon synthetic data and filtered publicly available websites - with a focus on high-quality, reasoning dense data. The model belongs to the Phi-4. Microsoft: Phi 4 Mini Instruct is listed as a large language or generative AI model associated with microsoft, with the model identifier microsoft/phi-4-mini-instruct. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
OpenAI: GPT-5 Image Mini
LLM Models
GPT-5 Image Mini combines OpenAI's advanced language capabilities, powered by [GPT-5 Mini](https://openrouter.ai/openai/gpt-5-mini), with GPT Image 1 Mini for efficient image generation. This natively multimodal model features superior instruction following, text. OpenAI: GPT-5 Image Mini is listed as a large language or generative AI model associated with openai, with the model identifier openai/gpt-5-image-mini. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Anthropic: Claude Haiku 4.5
LLM Models
Claude Haiku 4.5 is Anthropic’s fastest and most efficient model, delivering near-frontier intelligence at a fraction of the cost and latency of larger Claude models. Matching Claude Sonnet 4’s performance. Anthropic: Claude Haiku 4.5 is listed as a large language or generative AI model associated with anthropic, with the model identifier anthropic/claude-haiku-4.5. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Qwen: Qwen3 VL 8B Thinking
LLM Models
Qwen3-VL-8B-Thinking is the reasoning-optimized variant of the Qwen3-VL-8B multimodal model, designed for advanced visual and textual reasoning across complex scenes, documents, and temporal sequences. It integrates enhanced multimodal alignment and. Qwen: Qwen3 VL 8B Thinking is listed as a large language or generative AI model associated with qwen, with the model identifier qwen/qwen3-vl-8b-thinking. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Qwen: Qwen3 VL 8B Instruct
LLM Models
Qwen3-VL-8B-Instruct is a multimodal vision-language model from the Qwen3-VL series, built for high-fidelity understanding and reasoning across text, images, and video. It features improved multimodal fusion with Interleaved-MRoPE for long-horizon. Qwen: Qwen3 VL 8B Instruct is listed as a large language or generative AI model associated with qwen, with the model identifier qwen/qwen3-vl-8b-instruct. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
OpenAI: GPT-5 Image
LLM Models
[GPT-5](https://openrouter.ai/openai/gpt-5) Image combines OpenAI's GPT-5 model with state-of-the-art image generation capabilities. It offers major improvements in reasoning, code quality, and user experience while incorporating GPT Image 1's superior instruction following,. OpenAI: GPT-5 Image is listed as a large language or generative AI model associated with openai, with the model identifier openai/gpt-5-image. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
OpenAI: o3 Deep Research
LLM Models
o3-deep-research is OpenAI's advanced model for deep research, designed to tackle complex, multi-step research tasks. Note: This model always uses the 'web_search' tool which adds additional cost. OpenAI: o3 Deep Research is listed as a large language or generative AI model associated with openai, with the model identifier openai/o3-deep-research. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
OpenAI: o4 Mini Deep Research
LLM Models
o4-mini-deep-research is OpenAI's faster, more affordable deep research model—ideal for tackling complex, multi-step research tasks. Note: This model always uses the 'web_search' tool which adds additional cost. OpenAI: o4 Mini Deep Research is listed as a large language or generative AI model associated with openai, with the model identifier openai/o4-mini-deep-research. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Llama-3.3-Nemotron-Super-49B-v1.5 is a 49B-parameter, English-centric reasoning/chat model derived from Meta’s Llama-3.3-70B-Instruct with a 128K context. It’s post-trained for agentic workflows (RAG, tool calling) via SFT across math, code, science, and. NVIDIA: Llama 3.3 Nemotron Super 49B V1.5 is listed as a large language or generative AI model associated with nvidia, with the model identifier nvidia/llama-3.3-nemotron-super-49b-v1.5. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Gemini 2.5 Flash Image, a.k.a. "Nano Banana," is now generally available. It is a state of the art image generation model with contextual understanding. It is capable of image generation,. Google: Nano Banana (Gemini 2.5 Flash Image) is listed as a large language or generative AI model associated with google, with the model identifier google/gemini-2.5-flash-image. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Qwen: Qwen3 VL 30B A3B Thinking
LLM Models
Qwen3-VL-30B-A3B-Thinking is a multimodal model that unifies strong text generation with visual understanding for images and videos. Its Thinking variant enhances reasoning in STEM, math, and complex tasks. It excels. Qwen: Qwen3 VL 30B A3B Thinking is listed as a large language or generative AI model associated with qwen, with the model identifier qwen/qwen3-vl-30b-a3b-thinking. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Qwen: Qwen3 VL 30B A3B Instruct
LLM Models
Qwen3-VL-30B-A3B-Instruct is a multimodal model that unifies strong text generation with visual understanding for images and videos. Its Instruct variant optimizes instruction-following for general multimodal tasks. It excels in perception. Qwen: Qwen3 VL 30B A3B Instruct is listed as a large language or generative AI model associated with qwen, with the model identifier qwen/qwen3-vl-30b-a3b-instruct. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
OpenAI: GPT-5 Pro
LLM Models
GPT-5 Pro is OpenAI’s most advanced model, offering major improvements in reasoning, code quality, and user experience. It is optimized for complex tasks that require step-by-step reasoning, instruction following, and. OpenAI: GPT-5 Pro is listed as a large language or generative AI model associated with openai, with the model identifier openai/gpt-5-pro. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Z.ai: GLM 4.6
LLM Models
Compared with GLM-4.5, this generation brings several key improvements: Longer context window: The context window has been expanded from 128K to 200K tokens, enabling the model to handle more complex. Z.ai: GLM 4.6 is listed as a large language or generative AI model associated with z-ai, with the model identifier z-ai/glm-4.6. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Anthropic: Claude Sonnet 4.5
LLM Models
Claude Sonnet 4.5 is Anthropic’s most advanced Sonnet model to date, optimized for real-world agents and coding workflows. It delivers state-of-the-art performance on coding benchmarks such as SWE-bench Verified, with. Anthropic: Claude Sonnet 4.5 is listed as a large language or generative AI model associated with anthropic, with the model identifier anthropic/claude-sonnet-4.5. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
DeepSeek: DeepSeek V3.2 Exp
LLM Models
DeepSeek-V3.2-Exp is an experimental large language model released by DeepSeek as an intermediate step between V3.1 and future architectures. It introduces DeepSeek Sparse Attention (DSA), a fine-grained sparse attention mechanism. DeepSeek: DeepSeek V3.2 Exp is listed as a large language or generative AI model associated with deepseek, with the model identifier deepseek/deepseek-v3.2-exp. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
TheDrummer: Cydonia 24B V4.1
LLM Models
Uncensored and creative writing model based on Mistral Small 3.2 24B with good recall, prompt adherence, and intelligence. TheDrummer: Cydonia 24B V4.1 is listed as a large language or generative AI model associated with thedrummer, with the model identifier thedrummer/cydonia-24b-v4.1. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Relace: Relace Apply 3
LLM Models
Relace Apply 3 is a specialized code-patching LLM that merges AI-suggested edits straight into your source files. It can apply updates from GPT-4o, Claude, and others into your files at. Relace: Relace Apply 3 is listed as a large language or generative AI model associated with relace, with the model identifier relace/relace-apply-3. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Gemini 2.5 Flash-Lite is a lightweight reasoning model in the Gemini 2.5 family, optimized for ultra-low latency and cost efficiency. It offers improved throughput, faster token generation, and better performance. Google: Gemini 2.5 Flash Lite Preview 09-2025 is listed as a large language or generative AI model associated with google, with the model identifier google/gemini-2.5-flash-lite-preview-09-2025. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Qwen: Qwen3 VL 235B A22B Thinking
LLM Models
Qwen3-VL-235B-A22B Thinking is a multimodal model that unifies strong text generation with visual understanding across images and video. The Thinking model is optimized for multimodal reasoning in STEM and math. Qwen: Qwen3 VL 235B A22B Thinking is listed as a large language or generative AI model associated with qwen, with the model identifier qwen/qwen3-vl-235b-a22b-thinking. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Qwen: Qwen3 VL 235B A22B Instruct
LLM Models
Qwen3-VL-235B-A22B Instruct is an open-weight multimodal model that unifies strong text generation with visual understanding across images and video. The Instruct model targets general vision-language use (VQA, document parsing, chart/table. Qwen: Qwen3 VL 235B A22B Instruct is listed as a large language or generative AI model associated with qwen, with the model identifier qwen/qwen3-vl-235b-a22b-instruct. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Qwen: Qwen3 Max
LLM Models
Qwen3-Max is an updated release built on the Qwen3 series, offering major improvements in reasoning, instruction following, multilingual support, and long-tail knowledge coverage compared to the January 2025 version. It. Qwen: Qwen3 Max is listed as a large language or generative AI model associated with qwen, with the model identifier qwen/qwen3-max. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Qwen: Qwen3 Coder Plus
LLM Models
Qwen3 Coder Plus is Alibaba's proprietary version of the Open Source Qwen3 Coder 480B A35B. It is a powerful coding agent model specializing in autonomous programming via tool calling and. Qwen: Qwen3 Coder Plus is listed as a large language or generative AI model associated with qwen, with the model identifier qwen/qwen3-coder-plus. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
OpenAI: GPT-5 Codex
LLM Models
GPT-5-Codex is a specialized version of GPT-5 optimized for software engineering and coding workflows. It is designed for both interactive development sessions and long, independent execution of complex engineering tasks. OpenAI: GPT-5 Codex is listed as a large language or generative AI model associated with openai, with the model identifier openai/gpt-5-codex. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
DeepSeek: DeepSeek V3.1 Terminus
LLM Models
DeepSeek-V3.1 Terminus is an update to [DeepSeek V3.1](/deepseek/deepseek-chat-v3.1) that maintains the model's original capabilities while addressing issues reported by users, including language consistency and agent capabilities, further optimizing the model's. DeepSeek: DeepSeek V3.1 Terminus is listed as a large language or generative AI model associated with deepseek, with the model identifier deepseek/deepseek-v3.1-terminus. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Qwen: Qwen3 Coder Flash
LLM Models
Qwen3 Coder Flash is Alibaba's fast and cost efficient version of their proprietary Qwen3 Coder Plus. It is a powerful coding agent model specializing in autonomous programming via tool calling. Qwen: Qwen3 Coder Flash is listed as a large language or generative AI model associated with qwen, with the model identifier qwen/qwen3-coder-flash. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Qwen: Qwen3 Next 80B A3B Thinking
LLM Models
Qwen3-Next-80B-A3B-Thinking is a reasoning-first chat model in the Qwen3-Next line that outputs structured “thinking” traces by default. It’s designed for hard multi-step problems; math proofs, code synthesis/debugging, logic, and agentic. Qwen: Qwen3 Next 80B A3B Thinking is listed as a large language or generative AI model associated with qwen, with the model identifier qwen/qwen3-next-80b-a3b-thinking. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Qwen: Qwen3 Next 80B A3B Instruct (free)
LLM Models
Qwen3-Next-80B-A3B-Instruct is an instruction-tuned chat model in the Qwen3-Next series optimized for fast, stable responses without “thinking” traces. It targets complex tasks across reasoning, code generation, knowledge QA, and multilingual. Qwen: Qwen3 Next 80B A3B Instruct (free) is listed as a large language or generative AI model associated with qwen, with the model identifier qwen/qwen3-next-80b-a3b-instruct:free. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Qwen: Qwen3 Next 80B A3B Instruct
LLM Models
Qwen3-Next-80B-A3B-Instruct is an instruction-tuned chat model in the Qwen3-Next series optimized for fast, stable responses without “thinking” traces. It targets complex tasks across reasoning, code generation, knowledge QA, and multilingual. Qwen: Qwen3 Next 80B A3B Instruct is listed as a large language or generative AI model associated with qwen, with the model identifier qwen/qwen3-next-80b-a3b-instruct. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Qwen: Qwen Plus 0728 (thinking)
LLM Models
Qwen Plus 0728, based on the Qwen3 foundation model, is a 1 million context hybrid reasoning model with a balanced performance, speed, and cost combination. Qwen: Qwen Plus 0728 (thinking) is listed as a large language or generative AI model associated with qwen, with the model identifier qwen/qwen-plus-2025-07-28:thinking. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Qwen: Qwen Plus 0728
LLM Models
Qwen Plus 0728, based on the Qwen3 foundation model, is a 1 million context hybrid reasoning model with a balanced performance, speed, and cost combination. Qwen: Qwen Plus 0728 is listed as a large language or generative AI model associated with qwen, with the model identifier qwen/qwen-plus-2025-07-28. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
NVIDIA: Nemotron Nano 9B V2 (free)
LLM Models
NVIDIA-Nemotron-Nano-9B-v2 is a large language model (LLM) trained from scratch by NVIDIA, and designed as a unified model for both reasoning and non-reasoning tasks. It responds to user queries and. NVIDIA: Nemotron Nano 9B V2 (free) is listed as a large language or generative AI model associated with nvidia, with the model identifier nvidia/nemotron-nano-9b-v2:free. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
NVIDIA: Nemotron Nano 9B V2
LLM Models
NVIDIA-Nemotron-Nano-9B-v2 is a large language model (LLM) trained from scratch by NVIDIA, and designed as a unified model for both reasoning and non-reasoning tasks. It responds to user queries and. NVIDIA: Nemotron Nano 9B V2 is listed as a large language or generative AI model associated with nvidia, with the model identifier nvidia/nemotron-nano-9b-v2. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
MoonshotAI: Kimi K2 0905
LLM Models
Kimi K2 0905 is the September update of [Kimi K2 0711](moonshotai/kimi-k2). It is a large-scale Mixture-of-Experts (MoE) language model developed by Moonshot AI, featuring 1 trillion total parameters with 32. MoonshotAI: Kimi K2 0905 is listed as a large language or generative AI model associated with moonshotai, with the model identifier moonshotai/kimi-k2-0905. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Qwen: Qwen3 30B A3B Thinking 2507
LLM Models
Qwen3-30B-A3B-Thinking-2507 is a 30B parameter Mixture-of-Experts reasoning model optimized for complex tasks requiring extended multi-step thinking. The model is designed specifically for “thinking mode,” where internal reasoning traces are separated. Qwen: Qwen3 30B A3B Thinking 2507 is listed as a large language or generative AI model associated with qwen, with the model identifier qwen/qwen3-30b-a3b-thinking-2507. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Nous: Hermes 4 70B
LLM Models
Hermes 4 70B is a hybrid reasoning model from Nous Research, built on Meta-Llama-3.1-70B. It introduces the same hybrid mode as the larger 405B release, allowing the model to either. Nous: Hermes 4 70B is listed as a large language or generative AI model associated with nousresearch, with the model identifier nousresearch/hermes-4-70b. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Nous: Hermes 4 405B
LLM Models
Hermes 4 is a large-scale reasoning model built on Meta-Llama-3.1-405B and released by Nous Research. It introduces a hybrid reasoning mode, where the model can choose to deliberate internally with. Nous: Hermes 4 405B is listed as a large language or generative AI model associated with nousresearch, with the model identifier nousresearch/hermes-4-405b. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
DeepSeek: DeepSeek V3.1
LLM Models
DeepSeek-V3.1 is a large hybrid reasoning model (671B parameters, 37B active) that supports both thinking and non-thinking modes via prompt templates. It extends the DeepSeek-V3 base with a two-phase long-context. DeepSeek: DeepSeek V3.1 is listed as a large language or generative AI model associated with deepseek, with the model identifier deepseek/deepseek-chat-v3.1. It can be considered by developers, researchers, and AI application builders when evaluating models for programmatic AI workflows, experimentation, prototyping, model comparison, or integration into applications. The supplied description above captures the model-specific capabilities available in this dataset; exact context limits, supported input and output modalities, tool-calling behavior, structured-output support, pricing, rate limits, licensing, deployment options, and availability should be confirmed from the linked provider or model documentation because these technical details can vary between model versions and may change over time. For production use, developers should also evaluate output quality, latency, reliability, safety behavior, privacy requirements, and workload-specific performance against their own application requirements rather than assuming that one model configuration is suitable for every use case.
Frequently Asked Questions About LLM Models
What are LLM models and how do they work?
Large Language Models (LLMs) are AI models trained on vast amounts of text data to understand and generate human-like text. They use transformer architecture and can perform tasks like text generation, translation, summarization, and code completion.
Which LLM models are free to use?
Many LLM models are freely available including Llama 2, Mistral, Mixtral, Phi-2, and various models on Hugging Face. Our directory features only models with permissive licenses that allow commercial use.
How do I run LLM models locally?
You can run LLMs locally using tools like Ollama, LM Studio, or text-generation-webui. Most models require significant GPU memory (8GB+ VRAM recommended). Smaller models like Phi-2 and Llama 2 7B can run on consumer hardware.
What's the difference between open-source and proprietary LLMs?
Open-source LLMs like Llama 2 and Mistral have publicly available weights and can be run locally or fine-tuned. Proprietary models like GPT-4 are only accessible via API. Open-source models offer more control and privacy.
Can I use these LLM models for commercial projects?
Most models in our directory have permissive licenses (Apache 2.0, MIT) allowing commercial use. Always check the specific model's license on Hugging Face or the provider's website before using in production.