All LLM Models
Browse 1475 LLM models with VRAM requirements, quantization options, and hardware compatibility.
Understanding LLM VRAM Requirements
How much VRAM you need depends on the model size and quantization level. Quantization reduces the precision of model weights, trading small quality losses for significantly lower VRAM usage. For example, a 7B parameter model needs ~14 GB at FP16 but only ~4 GB at Q4_K_M quantization.
Model List
DeepSeek V4 Flash 0731 Reap 150B
puwaer · 150.1B · runs from 58.5 GB
1.5K
1DeepSeek V4 Flash 0731 Reap 150B is a 150.1B-parameter open language model from puwaer in the DeepSeek V4 family. It supports a context window of up to 1,048,576 tokens. See its VRAM requirements by quantization and which GPUs and Macs can run it locally below.
Hy MT2 7B
Tencent · 8.0B · runs from 2.8 GB
Hy-MT2-7B is Tencent's mid-size entry in the Hy-MT2 family of dedicated, "fast-thinking" multilingual translation models, alongside 1.8B and 30B-A3B (mixture-of-experts) siblings. Rather than being a general-purpose chat assistant, it is trained specifically for translation across 33 languages and to follow multilingual translation instructions, and the card reports it outperforming open models such as DeepSeek-V4-Pro and Kimi K2.6 in fast-thinking mode. Tencent also ships FP8 and GGUF quantized builds for on-device use with llama.cpp. At roughly 8 billion dense parameters, it fits on a single consumer GPU once quantized. Context length is 262,144 tokens. It is released under the Apache 2.0 license, permitting unrestricted commercial and research use, and was published in May 2026, alongside the open-sourced IFMTBench instruction-following translation benchmark.
DeepSeek R1 0528
DeepSeek · 684.5B · runs from 192.1 GB
DeepSeek R1 0528 is an updated release of the R1 reasoning model, incorporating improvements to training and inference that sharpen its performance on complex multi-step problems. It retains the same 684.5 billion parameter mixture-of-experts architecture as the original R1, with approximately 37 billion parameters active per forward pass. This revision addresses several edge cases where the original R1 struggled, delivering more consistent reasoning chains and fewer hallucinations on difficult math and coding tasks. Hardware requirements remain identical to the original R1, so users already set up to run the first version can swap in the 0528 weights with no changes to their infrastructure.
SmolLM3 3B
Hugging Face · 3.1B · runs from 1.3 GB
SmolLM3 3B is Hugging Face's latest-generation compact language model, representing a significant step up from the SmolLM2 series. At 3 billion parameters, it delivers considerably stronger reasoning, instruction following, and general language understanding while maintaining modest hardware requirements that keep it accessible on most consumer GPUs. This model benefits from improved training data, architectural refinements, and lessons learned from previous SmolLM generations. It is well positioned for local chatbot applications, coding assistance, and content generation tasks where you want strong performance without dedicating the resources required by 7B-class models.
Nex N2.5 Mini
nex-agi · 35.1B · runs from 10.0 GB
Nex N2.5 Mini is a 35.1B-parameter open language model from nex-agi. It supports a context window of up to 262,144 tokens. See its VRAM requirements by quantization and which GPUs and Macs can run it locally below.
DeepSeek R1 Distill Qwen 14B
DeepSeek · 14.8B · runs from 5.1 GB
DeepSeek R1 Distill Qwen 14B sits in a sweet spot between the smaller 7B distill and the more demanding 32B version, offering strong reasoning performance at 14.8 billion parameters on the Qwen 2.5 architecture. It captures a meaningful share of the full R1's chain-of-thought capabilities while keeping resource requirements within the range of mainstream consumer GPUs. Quantized to 4-bit, it fits comfortably on GPUs with 12 GB of VRAM, delivering reliable step-by-step reasoning for math, logic, and analytical problems.
DeepSeek V4 Flash Vision Exp
DeepSeek · 304.6B · runs from 84.1 GB
DeepSeek V4 Flash Vision Exp is an experimental multimodal model from DeepSeek, a mixture-of-experts design with roughly 304.6 billion total parameters, only a small fraction of which are active per token. It extends the text-only DeepSeek V4 Flash line with image understanding, built for agentic tasks, reasoning, and tool use. Only the active parameters are computed per token, keeping decoding fast, but all weights must fit in memory, making this a server-class model best run across multiple GPUs or a hosted endpoint. The model supports a 1,048,576 token context window, suited to very long documents or codebases. It is released under the MIT license, one of the most permissive open licenses, with no restrictions on commercial use. Published in August 2026 as an experimental release, it brings image input to the Flash line.
MiniCPM O 4 5
OpenBMB · 9.4B · runs from 4.6 GB
MiniCPM-o 4.5 is OpenBMB's roughly 9.4-billion-parameter omni-modal model, successor to MiniCPM-o 2.6, built from a SigLip2 vision encoder, a Whisper-medium audio encoder, a CosyVoice2 speech decoder, and a Qwen3-8B language backbone. It processes images, video, and audio and can generate spoken output, adding full-duplex live streaming so it can see, listen, and speak at once, plus a proactive mode that lets it comment unprompted, and bundled instruct and thinking modes. At this size it needs a high-end consumer GPU once quantized. Context length is 40,960 tokens. It is released under the Apache 2.0 license, permitting unrestricted commercial and research use, and was published in February 2026. Its defining upgrade over 2.6 is full-duplex interaction: earlier omni models process input and generate output in alternating turns, while 4.5 streams both directions at once without blocking.
Qwen2.5 Omni 7B
Alibaba · 10.7B · runs from 3.3 GB
Qwen2.5-Omni-7B is Alibaba's 10.7-billion-parameter flagship in the Qwen2.5-Omni family, an end-to-end model that perceives text, images, audio, and video and generates both text and natural streaming speech in response. It shares the family's Thinker-Talker design and TMRoPE positional scheme for synchronizing audio and video timestamps, tuned for low-latency, real-time conversation rather than turn-based chat alone. At just under 11 billion parameters, it needs a single mainstream-to-high-end consumer GPU once quantized. Its language backbone carries a 32,768 token context window. It is released under Alibaba's Qwen Research License, a non-commercial license limited to research and evaluation rather than the Apache 2.0 used for Qwen's text-only models. Published in March 2025, it was the first Omni release, with the smaller 3B variant following about a month later.
MiMo V2.6 Distill Qwen 9B
Xiaomi · 9.4B · runs from 3.4 GB
MiMo V2.6 Distill Qwen 9B is Xiaomi's 9.4-billion-parameter agentic model, built by fine-tuning Alibaba's Qwen3.5-9B on data generated by Xiaomi's larger MiMo-V2.6-Pro and MiMo-V2.6-Flash models. It handles vision, code, and function-calling tasks, and is positioned as the practical local-hardware option next to Xiaomi's much larger frontier MiMo-V2.6 models, which are designed for hosted use. At under 10 billion parameters, it runs well on a single mainstream consumer GPU once quantized. The model offers a 262K token context window, suited to long agent traces or multi-file coding sessions. Published in September 2026, it is one of the newest releases in Xiaomi's MiMo line, distilling the larger models' agentic and coding behaviour into a size that fits ordinary hardware.
Nanbeige4.1 3B
Nanbeige · 3.9B · runs from 2.1 GB
Nanbeige4.1 3B is a compact chat model from Nanbeige, a Chinese AI startup focused on building efficient small-scale language models. At just under 4 billion parameters, it is designed to run on virtually any modern GPU or even on CPU, making it one of the more accessible options for users with limited hardware. Despite its small size, it handles basic conversation, simple reasoning, and Chinese-English bilingual tasks, serving as a practical entry point for local LLM experimentation.
K2 Horizon MoVA 36B A4B
IFM · 37.4B · runs from 10.8 GB
K2 Horizon MoVA 36B A4B is a 37.4B-parameter open language model from IFM. It supports a context window of up to 524,288 tokens. See its VRAM requirements by quantization and which GPUs and Macs can run it locally below.
Qwen3.8 Whittle MoE 27B A17.8B
logic65 · 26.9B · runs from 12.2 GB
Qwen3.8 Whittle MoE 27B A17.8B is a 26.9B-parameter open language model from logic65 in the Qwen 3.8 family. It supports a context window of up to 262,144 tokens. See its VRAM requirements by quantization and which GPUs and Macs can run it locally below.
Kimi K2 Instruct
Moonshot AI · 1026.4B · runs from 286.1 GB
Kimi K2 Instruct is Moonshot AI's massive Mixture-of-Experts model, weighing in at over one trillion total parameters. It represents one of the largest open-weight models available, delivering frontier-class performance across reasoning, coding, and multilingual tasks through its sparse MoE architecture that activates only a fraction of its full parameter count per token. Running Kimi K2 locally is an extreme undertaking, requiring professional multi-GPU setups with hundreds of gigabytes of combined VRAM even at aggressive quantization. This model is best suited for research labs, enterprise deployments, or enthusiasts with access to server-grade hardware who want to explore trillion-parameter-scale inference.
Qwen3.5 27B Claude 4.6 Opus Reasoning Distilled
Jackrong · 27.8B · runs from 8.4 GB
The full-precision version of Jackrong's Qwen3.5 27B reasoning distillation from Claude 4.6 Opus. With 27.8 billion parameters in unquantized form, this model preserves the maximum quality from the distillation process but requires significantly more VRAM, typically 56 GB or more in BF16. It is primarily intended for users with professional-grade GPUs or multi-GPU setups. This variant is ideal for further fine-tuning, experimentation, or running at full fidelity when hardware allows. Most users looking to run the model locally for inference should consider the GGUF-quantized version instead, which offers a much better tradeoff between quality and resource usage.
Qwen3 235B A22B Thinking 2507
Alibaba · 235.1B · runs from 71.0 GB
Qwen3 235B A22B Thinking 2507 is the reasoning and chain-of-thought variant of Alibaba's largest Qwen3 mixture-of-experts model, updated in July 2025. With 235 billion total parameters and about 22 billion active per forward pass, it represents the pinnacle of Qwen3's reasoning capabilities. This model excels at complex multi-step problems, mathematical reasoning, code analysis, and tasks requiring deep logical thinking. It demands serious hardware to run locally, but for users with multi-GPU setups, it offers reasoning performance that rivals the best proprietary models while keeping all computation on your own machines.
DeepSeek R1 Distill Llama 8B
DeepSeek · 8.0B · runs from 2.8 GB
DeepSeek R1 Distill Llama 8B brings R1's reinforcement-learned reasoning capabilities to the widely supported Llama 3.1 8B architecture. By distilling the full 684.5B R1 model's reasoning patterns into this 8 billion parameter dense model, DeepSeek created a version that benefits from the extensive Llama ecosystem of tools, quantizations, and inference engines. For users who prefer the Llama architecture or already have tooling built around it, this model offers a plug-and-play path to chain-of-thought reasoning. Its hardware requirements are very approachable, running well on consumer GPUs with 8 GB or more of VRAM at common quantization levels.
Mistral 7B Instruct v0.2
Mistral AI · 7.2B · runs from 3.6 GB
Mistral 7B Instruct v0.2 is a 7.2-billion-parameter instruction-tuned language model from Mistral AI, built for general chat, question answering, and instruction following. It refines the original Mistral 7B with better adherence to complex prompts and improved handling of longer inputs. Its compact size suits local deployment on modern consumer GPUs, running comfortably on mainstream hardware once quantized. The model supports a 32,768 token context window, a result of raising the RoPE base frequency used for positional encoding, which improved handling of longer inputs versus the original v0.1. It is released under the Apache 2.0 license, allowing unrestricted commercial and research use. Published in December 2023, it became one of the most widely adopted 7B open-weight chat models, still supported by tools such as llama.cpp, vLLM, and Ollama.
MiniCPM V 4 5
OpenBMB · 8.7B · runs from 4.3 GB
MiniCPM-V 4.5 is an 8.7-billion-parameter vision-language model from OpenBMB, combining a Qwen3-8B language backbone with a SigLIP2 vision encoder. It is designed to run efficiently on modest hardware, including phones and laptops, while handling image, multi-image, and video understanding alongside text chat; a unified resampler compresses video frames heavily so longer clips don't demand proportionally more compute. At under 9 billion parameters, it fits comfortably on a single mainstream consumer GPU once quantized. It supports a 40,960 token context window, sufficient for long documents or multi-turn multimodal conversations. It is released under the Apache 2.0 license, allowing unrestricted commercial and research use. Published in August 2025, it distinguishes itself with a switchable "fast" and "deep" thinking mode, trading response speed for more deliberate reasoning on harder problems.
Kimi Linear 48B A3B Instruct
Moonshot AI · 49.1B · runs from 14.3 GB
Kimi Linear 48B A3B Instruct is a 49.1B-parameter open language model from Moonshot AI in the Kimi family. See its VRAM requirements by quantization and which GPUs and Macs can run it locally below.
Gemma 3n E2B IT
Google · 5.4B · runs from 1.6 GB
Gemma 3n E2B is the smaller of Google's two Gemma 3n instruction-tuned models, designed for phones, laptops and other on-device use. "E2B" stands for an effective size of about 2 billion parameters: the raw checkpoint holds around 5.4 billion, but techniques such as per-layer embeddings let it run with a memory footprint closer to a 2B model. It accepts text, image and audio input and generates text. The model has a 32K-token context window and is released under Google's Gemma terms of use. At 4-bit quantization it needs only a few gigabytes of memory, so it runs on almost any modern GPU, laptop or recent phone.
Qwen2.5 Math 7B Instruct
Alibaba · 7.6B · runs from 3.0 GB
Qwen2.5 Math 7B Instruct is a 7.6B-parameter open language model from Alibaba in the Qwen 2.5 family. It supports a context window of up to 4,096 tokens. See its VRAM requirements by quantization and which GPUs and Macs can run it locally below.
Qwen3.6 27B AEON Ultimate Uncensored
AEON-7 · 27.4B · runs from 12.4 GB
Qwen3.6 27B AEON Ultimate Uncensored is a 27.4B-parameter open language model from AEON-7 in the Qwen 3.6 family. It supports a context window of up to 262,144 tokens. See its VRAM requirements by quantization and which GPUs and Macs can run it locally below.
Qwen2.5 Coder 7B Instruct Abliterated
huihui-ai · 7.6B · runs from 3.0 GB
Qwen2.5 Coder 7B Instruct Abliterated is a 7.6B-parameter open language model from huihui-ai in the Qwen 2.5 family. It supports a context window of up to 32,768 tokens. See its VRAM requirements by quantization and which GPUs and Macs can run it locally below.
Mistral 7B Instruct v0.1
Mistral AI · 7.2B · runs from 3.6 GB
Mistral 7B Instruct v0.1 was the first instruction-tuned variant of the original Mistral 7B, fine-tuned for conversational and instruction-following tasks. While it has since been superseded by v0.2 and v0.3, it remains a solid lightweight chat model and an important milestone in the open-weight model ecosystem. Its hardware requirements are identical to the base Mistral 7B, running smoothly on GPUs with as little as 6 GB of VRAM when quantized. Users seeking the best Mistral 7B experience should generally prefer the newer v0.3 release, but v0.1 is still useful for reproducibility and benchmarking purposes.
DeepSeek V4 Pro
DeepSeek · 1598.8B · runs from 679.8 GB
DeepSeek-V4-Pro is a preview release in DeepSeek's V4 series of Mixture-of-Experts language models, built for highly efficient million-token-context reasoning. It totals roughly 1.6 trillion parameters with 49 billion activated per token, pre-trained on more than 32 trillion tokens and post-trained through supervised fine-tuning, reinforcement learning, and on-policy distillation. This instruct release supports three switchable reasoning-effort modes, from fast everyday answers up to an extended "Think Max" mode for harder problems, plus stronger tool-calling and agentic behavior. It natively handles a one-million-token context window and is released under the MIT license. At this scale, running it locally means server-class or multi-GPU hardware, with 4-bit quantization still needing roughly 900GB of memory; most people will reach it through a hosted endpoint instead.
DeepSeek V4.1 Flash
DeepSeek · 763.2B · runs from 324.7 GB
DeepSeek-V4.1-Flash is DeepSeek's multimodal Mixture-of-Experts model, processing images and text together and generating text autoregressively. It uses a Causal Encoder-Decoder architecture with roughly 763 billion total parameters, but activates only about 8 billion during prefill and 16 billion during decode, keeping input-heavy agentic workloads efficient. The model supports a continuously adjustable reasoning-effort setting that trades inference cost for accuracy, and was trained with aggressive KV-cache compression aimed at long-context, tool-using agents. Context reaches up to one million tokens, and it is released under the MIT license. Given its size, local use calls for multi-GPU or server-class hardware, with roughly 439GB of memory needed even at 4-bit; most users will rely on a hosted API instead.
Gemma 4 12B IT Heretic
igorls · 12.0B · runs from 6.1 GB
Gemma 4 12B IT Heretic is a 12.0B-parameter open language model from igorls in the Gemma 4 family. It supports a context window of up to 131,072 tokens. See its VRAM requirements by quantization and which GPUs and Macs can run it locally below.
Huihui Qwen3 Coder 30B A3B Instruct Abliterated
huihui-ai · 30.5B · runs from 8.8 GB
Huihui Qwen3 Coder 30B A3B Instruct Abliterated is a 30.5B-parameter open language model from huihui-ai in the Qwen 3 family. It supports a context window of up to 262,144 tokens. See its VRAM requirements by quantization and which GPUs and Macs can run it locally below.
Gemma 4 E2B IT Qat Mobile Transformers
Google · 2.3B · runs from 1.4 GB
Gemma 4 E2B IT Qat Mobile Transformers is a 2.3B-parameter open language model from Google in the Gemma 4 family. It supports a context window of up to 131,072 tokens. See its VRAM requirements by quantization and which GPUs and Macs can run it locally below.