# ModelFit — Can I Run This LLM?

> Measured GGUF file sizes and real VRAM/RAM requirements, updated daily from the Hugging Face API.
Every number below is the actual published file size plus KV-cache and runtime overhead — not a formula estimate.

## Models

- [Qwen3-Coder-30B-A3B-Instruct](https://modelfit-eight.vercel.app/unsloth-qwen3-coder-30b-a3b-instruct-gguf/index.md): recommended UD-IQ1_M ~13.1 GB, smallest ~11.1 GB
- [Qwen3.8-27B](https://modelfit-eight.vercel.app/unsloth-qwen3-8-27b-gguf/index.md): recommended UD-Q2_K_XL ~13.0 GB, smallest ~9.3 GB
- [Ornith-1.5-9B](https://modelfit-eight.vercel.app/ornith-ai-ornith-1-5-9b-gguf/index.md): recommended Q4_K_M ~8.4 GB, smallest ~8.4 GB
- [Ornith-1.5-35B-A3B](https://modelfit-eight.vercel.app/ornith-ai-ornith-1-5-35b-a3b-gguf/index.md): recommended Q4_K_M ~27.6 GB, smallest ~27.6 GB
- [Ornith-1.0-9B](https://modelfit-eight.vercel.app/ornith-ai-ornith-1-0-9b-gguf/index.md): recommended Q4_K_M ~8.3 GB, smallest ~8.3 GB
- [Huihui-Qwen3.8-27B-abliterated](https://modelfit-eight.vercel.app/huihui-ai-huihui-qwen3-8-27b-abliterated-gguf/index.md): recommended UD-Q2_K_XL ~13.5 GB, smallest ~9.4 GB
- [Qwen3.8-27B-Uncensored](https://modelfit-eight.vercel.app/jonathancoletti-qwen3-8-27b-uncensored-gguf/index.md): recommended IQ2_M ~26.5 GB, smallest ~3.5 GB
- [Qwen3.8-27B-Uncensored-HauhauCS-Aggressive-MTP](https://modelfit-eight.vercel.app/hauhaucs-qwen3-8-27b-uncensored-hauhaucs-aggressive-mtp-gguf/index.md): recommended IQ2_M ~13.9 GB, smallest ~13.9 GB
- [Ornith-1.0-35B](https://modelfit-eight.vercel.app/ornith-ai-ornith-1-0-35b-gguf/index.md): recommended Q4_K_M ~26.9 GB, smallest ~26.9 GB
- [deepseek-v4](https://modelfit-eight.vercel.app/antirez-deepseek-v4-gguf/index.md): recommended Q2KDown-AProjQ8-SExpQ8-OutQ8 ~118.6 GB, smallest ~118.6 GB
- [Ternary-Bonsai-2-27B](https://modelfit-eight.vercel.app/prism-ml-ternary-bonsai-2-27b-gguf/index.md): recommended BF16 ~2.6 GB, smallest ~2.3 GB
- [Qwen3.8-27B](https://modelfit-eight.vercel.app/lmstudio-community-qwen3-8-27b-gguf/index.md): recommended Q6_K ~28.4 GB, smallest ~21.7 GB
- [Gemma-4-E4B-Uncensored-HauhauCS-Aggressive](https://modelfit-eight.vercel.app/hauhaucs-gemma-4-e4b-uncensored-hauhaucs-aggressive/index.md): recommended Q2_K_P ~6.8 GB, smallest ~6.8 GB
- [nemotron-3.5-asr-streaming-0.6b](https://modelfit-eight.vercel.app/handy-computer-nemotron-3-5-asr-streaming-0-6b-gguf/index.md): recommended Q4_K_M ~2.1 GB, smallest ~2.1 GB
- [Qwen3.8-27B-Heretic-Abliterated-Uncensored](https://modelfit-eight.vercel.app/0bserverx-qwen3-8-27b-heretic-abliterated-uncensored-gguf/index.md): recommended Q5_K_M ~24.6 GB, smallest ~13.8 GB
- [Qwen3.8-Flash-Next](https://modelfit-eight.vercel.app/unsloth-qwen3-8-flash-next-gguf/index.md): recommended UD-IQ1_S ~88.6 GB, smallest ~88.6 GB
- [parakeet-unified-en-0.6b](https://modelfit-eight.vercel.app/handy-computer-parakeet-unified-en-0-6b-gguf/index.md): recommended Q4_K_M ~2.1 GB, smallest ~2.1 GB
- [Qwen3.5-9B-The-Defiant-Fable-Uncensored-Heretic-NEO-IMATRIX-MAX-MTP](https://modelfit-eight.vercel.app/davidau-qwen3-5-9b-the-defiant-fable-uncensored-heretic-neo-imatrix-max-mtp-gguf/index.md): recommended IQ2_M ~13.4 GB, smallest ~13.4 GB
- [Qwen3.5-9B](https://modelfit-eight.vercel.app/unsloth-qwen3-5-9b-gguf/index.md): recommended UD-IQ2_XXS ~5.3 GB, smallest ~5.3 GB
- [gemma-4-E4B-it](https://modelfit-eight.vercel.app/ggml-org-gemma-4-e4b-it-gguf/index.md): recommended Q8_0 ~11.1 GB, smallest ~7.0 GB
- [Ornith-1.5-397B](https://modelfit-eight.vercel.app/ornith-ai-ornith-1-5-397b-gguf/index.md): recommended Q4_K_M ~294.7 GB, smallest ~294.7 GB
- [Inkling-Small](https://modelfit-eight.vercel.app/unsloth-inkling-small-gguf/index.md): recommended UD-IQ1_M ~96.1 GB, smallest ~91.2 GB
- [Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NEO-CODER-MAX-MTP](https://modelfit-eight.vercel.app/davidau-qwen3-8-27b-turbo-fable-cold-fusion-735-882-heretic-uncensored-neo-coder-max-mtp-gguf/index.md): recommended IQ2_M ~30.1 GB, smallest ~30.1 GB
- [Qwen3.6-35B-A3B](https://modelfit-eight.vercel.app/unsloth-qwen3-6-35b-a3b-gguf/index.md): recommended UD-IQ2_XXS ~14.4 GB, smallest ~13.6 GB
- [Qwen3.8-27B-OBLITERATED](https://modelfit-eight.vercel.app/obliteratus-qwen3-8-27b-obliterated/index.md): recommended Q2_K ~14.0 GB, smallest ~14.0 GB
- [Qwen3.8-27B-GSQ-RCO](https://modelfit-eight.vercel.app/ista-daslab-qwen3-8-27b-gsq-rco-gguf/index.md): recommended IQ2_XS ~22.1 GB, smallest ~22.1 GB
- [LFM2.5-2.6B](https://modelfit-eight.vercel.app/liquidai-lfm2-5-2-6b-gguf/index.md): recommended Q4_K_M ~3.5 GB, smallest ~3.5 GB
- [Qwen3.8-27B](https://modelfit-eight.vercel.app/ggml-org-qwen3-8-27b-gguf/index.md): recommended Q8_0 ~38.3 GB, smallest ~24.3 GB
- [Qwen3.6-27B-MTP](https://modelfit-eight.vercel.app/unsloth-qwen3-6-27b-mtp-gguf/index.md): recommended UD-IQ2_XXS ~13.0 GB, smallest ~13.0 GB
- [Qwen3.6-27B](https://modelfit-eight.vercel.app/unsloth-qwen3-6-27b-gguf/index.md): recommended UD-IQ2_XXS ~12.8 GB, smallest ~12.8 GB
- [Qwen3.6-35B-A3B-MTP](https://modelfit-eight.vercel.app/unsloth-qwen3-6-35b-a3b-mtp-gguf/index.md): recommended UD-IQ2_XXS ~15.7 GB, smallest ~15.1 GB
- [MiniMax-H3](https://modelfit-eight.vercel.app/unsloth-minimax-h3-gguf/index.md): recommended UD-Q3_K_XL ~13.0 GB, smallest ~11.2 GB
- [endless-frontier_BigBang-v1](https://modelfit-eight.vercel.app/bartowski-endless-frontier-bigbang-v1-gguf/index.md): recommended IQ2_XXS ~13.8 GB, smallest ~13.8 GB
- [glm-4-9b-chat-IMat](https://modelfit-eight.vercel.app/legraphista-glm-4-9b-chat-imat-gguf/index.md): recommended IQ1_S ~5.2 GB, smallest ~5.2 GB
- [cohere-transcribe-03-2026](https://modelfit-eight.vercel.app/handy-computer-cohere-transcribe-03-2026-gguf/index.md): recommended Q4_K_M ~3.4 GB, smallest ~3.4 GB
- [Tiel-Coder-35B-A3B-GGUF-MTP](https://modelfit-eight.vercel.app/peculiar-ragdoll-tiel-coder-35b-a3b-gguf-mtp/index.md): recommended UD-Q2_K_XL ~16.7 GB, smallest ~16.7 GB
- [Qwen3.5-4B](https://modelfit-eight.vercel.app/unsloth-qwen3-5-4b-gguf/index.md): recommended UD-IQ2_XXS ~3.3 GB, smallest ~3.3 GB
- [gemma-4-12b-it](https://modelfit-eight.vercel.app/unsloth-gemma-4-12b-it-gguf/index.md): recommended UD-IQ2_M ~6.6 GB, smallest ~6.6 GB
- [Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive](https://modelfit-eight.vercel.app/hauhaucs-qwen3-6-35b-a3b-uncensored-hauhaucs-aggressive/index.md): recommended IQ2_M ~15.5 GB, smallest ~15.5 GB
- [XYZAILab_XYZ-Aquila-mini](https://modelfit-eight.vercel.app/bartowski-xyzailab-xyz-aquila-mini-gguf/index.md): recommended IQ2_XXS ~13.2 GB, smallest ~13.2 GB
- [POCKET-35B](https://modelfit-eight.vercel.app/final-bench-pocket-35b-gguf/index.md): recommended IQ1_M ~11.4 GB, smallest ~11.4 GB
- [Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-NEO-MAX-MTP](https://modelfit-eight.vercel.app/davidau-qwen3-6-27b-fable-fusion-711-uncensored-heretic-nm-dau-neo-max-mtp-gguf/index.md): recommended IQ2_M ~30.1 GB, smallest ~30.1 GB
- [Qwen3.8-4B-Distill](https://modelfit-eight.vercel.app/empero-ai-qwen3-8-4b-distill-gguf/index.md): recommended Q4_K_M ~4.8 GB, smallest ~4.8 GB
- [Qwen3.8-2B-Distill](https://modelfit-eight.vercel.app/empero-ai-qwen3-8-2b-distill-gguf/index.md): recommended Q4_K_M ~3.1 GB, smallest ~3.1 GB
- [gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2](https://modelfit-eight.vercel.app/yuxinlu1-gemma-4-12b-agentic-fable5-composer2-5-v2-3-5x-tau2-gguf/index.md): recommended Q3_K_M ~8.8 GB, smallest ~8.8 GB
- [Ternary-Bonsai-27B](https://modelfit-eight.vercel.app/prism-ml-ternary-bonsai-27b-gguf/index.md): recommended Q4_1 ~3.8 GB, smallest ~2.3 GB
- [GLM-5.3-Flash](https://modelfit-eight.vercel.app/unsloth-glm-5-3-flash-gguf/index.md): recommended UD-IQ1_S ~113.2 GB, smallest ~113.2 GB
- [Parable-Qwen3-8B-Claude-Fable-5](https://modelfit-eight.vercel.app/ankitai-parable-qwen3-8b-claude-fable-5-gguf/index.md): recommended Q4_K_M ~7.5 GB, smallest ~7.5 GB
- [Huihui-DeepSeek-V4-Flash-0731-abliterated](https://modelfit-eight.vercel.app/huihui-ai-huihui-deepseek-v4-flash-0731-abliterated-gguf/index.md): recommended Q2-0731 ~105.6 GB, smallest ~105.6 GB
- [GLM-5.3](https://modelfit-eight.vercel.app/unsloth-glm-5-3-gguf/index.md): recommended UD-IQ1_S ~261.6 GB, smallest ~261.6 GB
- [Kimi-K3](https://modelfit-eight.vercel.app/unsloth-kimi-k3-gguf/index.md): recommended UD-Q1_0 ~561.1 GB, smallest ~561.1 GB
- [Parable-Qwen3-4B-Claude-Fable-5](https://modelfit-eight.vercel.app/ankitai-parable-qwen3-4b-claude-fable-5-gguf/index.md): recommended Q4_K_M ~4.5 GB, smallest ~4.5 GB
- [Qwen3-4B](https://modelfit-eight.vercel.app/unsloth-qwen3-4b-gguf/index.md): recommended UD-IQ2_XXS ~3.5 GB, smallest ~3.3 GB
- [Tiel-Coder-35B-A3B](https://modelfit-eight.vercel.app/peculiar-ragdoll-tiel-coder-35b-a3b-gguf/index.md): recommended UD-Q2_K_XL ~16.2 GB, smallest ~16.2 GB
- [Qwen3-4B](https://modelfit-eight.vercel.app/qwen-qwen3-4b-gguf/index.md): recommended Q4_K_M ~4.5 GB, smallest ~4.5 GB
- [Qwen3.6-14B-A3B-FableVibes](https://modelfit-eight.vercel.app/tvall43-qwen3-6-14b-a3b-fablevibes-gguf/index.md): recommended Q2_K ~7.9 GB, smallest ~2.6 GB
- [Qwen3-8B](https://modelfit-eight.vercel.app/qwen-qwen3-8b-gguf/index.md): recommended Q4_K_M ~7.5 GB, smallest ~7.5 GB
- [Qwen3.5-9B-DeepSeek-V4-Flash](https://modelfit-eight.vercel.app/jackrong-qwen3-5-9b-deepseek-v4-flash-gguf/index.md): recommended Q3_K_M ~7.0 GB, smallest ~7.0 GB
- [Parable-Granite-4.1-3B-Claude-Fable-5](https://modelfit-eight.vercel.app/ankitai-parable-granite-4-1-3b-claude-fable-5-gguf/index.md): recommended Q4_K_M ~4.0 GB, smallest ~4.0 GB
- [Kimi-K2.7-Code](https://modelfit-eight.vercel.app/unsloth-kimi-k2-7-code-gguf/index.md): recommended UD-IQ1_M ~366.2 GB, smallest ~366.2 GB
- [Parable-Granite-4.1-8B-Claude-Fable-5](https://modelfit-eight.vercel.app/ankitai-parable-granite-4-1-8b-claude-fable-5-gguf/index.md): recommended Q4_K_M ~7.6 GB, smallest ~7.6 GB
- [granite-4.2-30b](https://modelfit-eight.vercel.app/ibm-granite-granite-4-2-30b-gguf/index.md): recommended Q2_K ~14.5 GB, smallest ~14.5 GB
- [GLM-5.2](https://modelfit-eight.vercel.app/unsloth-glm-5-2-gguf/index.md): recommended UD-IQ1_S ~261.6 GB, smallest ~261.6 GB
- [granite-4.2-8b](https://modelfit-eight.vercel.app/ibm-granite-granite-4-2-8b-gguf/index.md): recommended Q2_K ~5.6 GB, smallest ~5.6 GB
- [Meta-Llama-3.1-8B-Instruct](https://modelfit-eight.vercel.app/bartowski-meta-llama-3-1-8b-instruct-gguf/index.md): recommended IQ2_M ~5.0 GB, smallest ~5.0 GB
- [granite-4.2-3b](https://modelfit-eight.vercel.app/ibm-granite-granite-4-2-3b-gguf/index.md): recommended Q2_K ~3.2 GB, smallest ~3.2 GB
- [Kimi-K2.6](https://modelfit-eight.vercel.app/unsloth-kimi-k2-6-gguf/index.md): recommended UD-Q2_K_XL ~410.1 GB, smallest ~410.1 GB
- [Qwen3.5-9B-GLM5.1-Distill-v1](https://modelfit-eight.vercel.app/jackrong-qwen3-5-9b-glm5-1-distill-v1-gguf/index.md): recommended Q3_K_M ~7.0 GB, smallest ~7.0 GB
- [DeepSeek-V4-Flash-0731](https://modelfit-eight.vercel.app/unsloth-deepseek-v4-flash-0731-gguf/index.md): recommended UD-IQ1_S ~100.5 GB, smallest ~14.6 GB
- [Llama-3.2-3B-Instruct](https://modelfit-eight.vercel.app/unsloth-llama-3-2-3b-instruct-gguf/index.md): recommended UD-IQ2_XXS ~3.0 GB, smallest ~2.9 GB
- [Qwen3-14B](https://modelfit-eight.vercel.app/qwen-qwen3-14b-gguf/index.md): recommended Q4_K_M ~12.3 GB, smallest ~12.3 GB
- [MiniCPM5-2B](https://modelfit-eight.vercel.app/openbmb-minicpm5-2b-gguf/index.md): recommended Q8_0 ~4.7 GB, smallest ~3.4 GB
- [Qwen3.8-27B-TWIN-TURBO-Fable-Cold-Fusion-709-L-Uncensored-NM-DAU-NEO-MTP](https://modelfit-eight.vercel.app/davidau-qwen3-8-27b-twin-turbo-fable-cold-fusion-709-l-uncensored-nm-dau-neo-mtp-gguf/index.md): recommended Q4_K_S ~41.5 GB, smallest ~41.5 GB
- [Qwen2.5-7B-Instruct](https://modelfit-eight.vercel.app/bartowski-qwen2-5-7b-instruct-gguf/index.md): recommended IQ2_M ~4.8 GB, smallest ~4.8 GB
- [Qwen2.5-VL-7B-Instruct](https://modelfit-eight.vercel.app/unsloth-qwen2-5-vl-7b-instruct-gguf/index.md): recommended UD-IQ2_XXS ~4.4 GB, smallest ~4.0 GB
- [Llama-3.2-3B-Instruct](https://modelfit-eight.vercel.app/maziyarpanahi-llama-3-2-3b-instruct-gguf/index.md): recommended IQ1_S ~2.5 GB, smallest ~2.5 GB
- [Qwen2.5-7B-Instruct](https://modelfit-eight.vercel.app/maziyarpanahi-qwen2-5-7b-instruct-gguf/index.md): recommended IQ1_S ~3.8 GB, smallest ~3.8 GB
- [Mistral-Small-Instruct-2409](https://modelfit-eight.vercel.app/maziyarpanahi-mistral-small-instruct-2409-gguf/index.md): recommended IQ1_S ~7.3 GB, smallest ~7.3 GB
- [Swift-Qwen3.8-27B](https://modelfit-eight.vercel.app/ukisai-swift-qwen3-8-27b-gguf/index.md): recommended IQ2_XXS ~12.4 GB, smallest ~12.4 GB
- [Llama-3.2-3B-Instruct](https://modelfit-eight.vercel.app/bartowski-llama-3-2-3b-instruct-gguf/index.md): recommended IQ3_M ~3.4 GB, smallest ~3.4 GB
- [Phi-4-mini-instruct](https://modelfit-eight.vercel.app/unsloth-phi-4-mini-instruct-gguf/index.md): recommended Q2_K ~3.5 GB, smallest ~3.5 GB
- [Llama-3.2-3B-Instruct](https://modelfit-eight.vercel.app/lmstudio-community-llama-3-2-3b-instruct-gguf/index.md): recommended Q3_K_L ~3.7 GB, smallest ~3.7 GB
- [Qwen_Qwen3-14B](https://modelfit-eight.vercel.app/bartowski-qwen-qwen3-14b-gguf/index.md): recommended IQ2_XS ~7.1 GB, smallest ~7.1 GB
- [Meta-Llama-3.1-8B-Instruct](https://modelfit-eight.vercel.app/lmstudio-community-meta-llama-3-1-8b-instruct-gguf/index.md): recommended Q3_K_L ~6.7 GB, smallest ~6.7 GB
- [gemma-3-4b-it](https://modelfit-eight.vercel.app/unsloth-gemma-3-4b-it-gguf/index.md): recommended UD-IQ2_XXS ~3.5 GB, smallest ~3.3 GB
- [gemma-3-4b-it](https://modelfit-eight.vercel.app/lmstudio-community-gemma-3-4b-it-gguf/index.md): recommended Q3_K_L ~4.2 GB, smallest ~4.2 GB
- [gemma-3-12b-it](https://modelfit-eight.vercel.app/unsloth-gemma-3-12b-it-gguf/index.md): recommended UD-IQ2_XXS ~7.8 GB, smallest ~7.3 GB
- [Qwen3-14B](https://modelfit-eight.vercel.app/lmstudio-community-qwen3-14b-gguf/index.md): recommended Q3_K_L ~11.0 GB, smallest ~11.0 GB
- [Llama-3.2-3B-Instruct-uncensored](https://modelfit-eight.vercel.app/bartowski-llama-3-2-3b-instruct-uncensored-gguf/index.md): recommended Q2_K ~3.3 GB, smallest ~3.3 GB
- [Qwen3.8-35B-A3B-Distill](https://modelfit-eight.vercel.app/empero-ai-qwen3-8-35b-a3b-distill-gguf/index.md): recommended IQ2_M ~16.6 GB, smallest ~16.6 GB
- [Ministral-3-14B-Reasoning-2512](https://modelfit-eight.vercel.app/lmstudio-community-ministral-3-14b-reasoning-2512-gguf/index.md): recommended Q6_K ~14.8 GB, smallest ~11.4 GB
- [Mistral-Small-3.2-24B-Instruct-2506](https://modelfit-eight.vercel.app/unsloth-mistral-small-3-2-24b-instruct-2506-gguf/index.md): recommended UD-IQ2_XXS ~9.6 GB, smallest ~8.2 GB
- [Ministral-3-3B-Instruct-2512](https://modelfit-eight.vercel.app/lmstudio-community-ministral-3-3b-instruct-2512-gguf/index.md): recommended Q6_K ~4.9 GB, smallest ~4.1 GB
- [google_gemma-3-4b-it](https://modelfit-eight.vercel.app/bartowski-google-gemma-3-4b-it-gguf/index.md): recommended IQ2_M ~3.3 GB, smallest ~3.3 GB
- [microsoft_Phi-4-mini-instruct](https://modelfit-eight.vercel.app/bartowski-microsoft-phi-4-mini-instruct-gguf/index.md): recommended IQ2_M ~3.3 GB, smallest ~3.3 GB
- [Kimi-K2-Instruct](https://modelfit-eight.vercel.app/unsloth-kimi-k2-instruct-gguf/index.md): recommended UD-IQ2_XXS ~396.0 GB, smallest ~293.8 GB
- [Qwen3.8-Flash-Next-GSQ-RCO](https://modelfit-eight.vercel.app/ista-daslab-qwen3-8-flash-next-gsq-rco-gguf/index.md): recommended IQ2_XS ~83.1 GB, smallest ~81.2 GB
- [Dolphin3.0-Llama3.1-8B](https://modelfit-eight.vercel.app/dphn-dolphin3-0-llama3-1-8b-gguf/index.md): recommended Q2_K ~5.3 GB, smallest ~5.3 GB
- [Ministral-3-14B-Instruct-2512](https://modelfit-eight.vercel.app/mistralai-ministral-3-14b-instruct-2512-gguf/index.md): recommended Q4_K_M ~11.4 GB, smallest ~11.4 GB
- [gemma-3-12b-it-qat](https://modelfit-eight.vercel.app/unsloth-gemma-3-12b-it-qat-gguf/index.md): recommended UD-IQ2_XXS ~7.8 GB, smallest ~7.3 GB
- [Cyber-Tiel-Coder-35B-A3B-GGUF-MTP](https://modelfit-eight.vercel.app/peculiar-ragdoll-cyber-tiel-coder-35b-a3b-gguf-mtp/index.md): recommended UD-Q2_K_XL ~16.7 GB, smallest ~16.7 GB
- [Qwen3.8-27B-TWIN-TURBO-Fable-Cold-Fusion-709-ULTRA-HERETIC-Uncensored-NM-DAU-NEO-MTP](https://modelfit-eight.vercel.app/davidau-qwen3-8-27b-twin-turbo-fable-cold-fusion-709-ultra-heretic-uncensored-nm-dau-neo-mtp-gguf/index.md): recommended IQ2_M ~16.0 GB, smallest ~16.0 GB
- [Qwen3.8-27B](https://modelfit-eight.vercel.app/byteshape-qwen3-8-27b-gguf/index.md): recommended IQ2_XXS ~12.1 GB, smallest ~12.1 GB
- [ukisai_Swift-Qwen3.8-27b](https://modelfit-eight.vercel.app/bartowski-ukisai-swift-qwen3-8-27b-gguf/index.md): recommended IQ2_XXS ~12.2 GB, smallest ~12.2 GB
- [Mistral-Small-3.1-24B-Instruct-2503](https://modelfit-eight.vercel.app/unsloth-mistral-small-3-1-24b-instruct-2503-gguf/index.md): recommended UD-IQ2_XXS ~9.1 GB, smallest ~7.9 GB
- [Kimi-K2.5](https://modelfit-eight.vercel.app/unsloth-kimi-k2-5-gguf/index.md): recommended UD-IQ2_XXS ~393.6 GB, smallest ~288.9 GB
- [Kimi-K2-Thinking](https://modelfit-eight.vercel.app/unsloth-kimi-k2-thinking-gguf/index.md): recommended UD-IQ1_M ~372.6 GB, smallest ~297.7 GB
- [Kimi-K3](https://modelfit-eight.vercel.app/atomicchat-kimi-k3-gguf/index.md): recommended IQ1_S ~709.0 GB, smallest ~709.0 GB
- [Sentie1.0-3B-Claude-Fable-5-GPT5.2-Sol-Kimi-K3-GLM-5.2](https://modelfit-eight.vercel.app/sentieai-sentie1-0-3b-claude-fable-5-gpt5-2-sol-kimi-k3-glm-5-2-gguf/index.md): recommended IQ3_M ~3.8 GB, smallest ~3.8 GB
- [Qwen-Image-2.1-Uncensored](https://modelfit-eight.vercel.app/abenzerps-qwen-image-2-1-uncensored-gguf/index.md): recommended Q4_0 ~6.4 GB, smallest ~6.4 GB
- [Best LLM for 8GB VRAM](https://modelfit-eight.vercel.app/best-llm-for-8gb-vram/index.md): top pick Ternary-Bonsai-2-27B at BF16
- [Best LLM for 12GB VRAM](https://modelfit-eight.vercel.app/best-llm-for-12gb-vram/index.md): top pick Qwen3.8-27B at UD-IQ2_XXS
- [Best LLM for 16GB VRAM](https://modelfit-eight.vercel.app/best-llm-for-16gb-vram/index.md): top pick Qwen3.8-27B-Heretic-Abliterated-Uncensored at Q2_K
- [Best LLM for 24GB VRAM](https://modelfit-eight.vercel.app/best-llm-for-24gb-vram/index.md): top pick Qwen3.8-27B-Heretic-Abliterated-Uncensored at Q4_K_S
- [Best LLM for 32GB VRAM](https://modelfit-eight.vercel.app/best-llm-for-32gb-vram/index.md): top pick Qwen3.8-27B-Heretic-Abliterated-Uncensored at Q6_K
- [Best LLM for 48GB VRAM](https://modelfit-eight.vercel.app/best-llm-for-48gb-vram/index.md): top pick Qwen3.8-27B-Heretic-Abliterated-Uncensored at IQ2_XXS
- [Best LLM for 64GB VRAM](https://modelfit-eight.vercel.app/best-llm-for-64gb-vram/index.md): top pick Qwen3.8-27B-Heretic-Abliterated-Uncensored at IQ3_XXS
- [Best LLM for 96GB VRAM](https://modelfit-eight.vercel.app/best-llm-for-96gb-vram/index.md): top pick Qwen3.8-27B-Heretic-Abliterated-Uncensored at IQ4_NL
- [Best LLM for RTX 5090](https://modelfit-eight.vercel.app/best-llm-for-rtx-5090/index.md): top pick Qwen3.8-27B-Heretic-Abliterated-Uncensored at Q6_K
- [Best LLM for RTX 4090](https://modelfit-eight.vercel.app/best-llm-for-rtx-4090/index.md): top pick Qwen3.8-27B-Heretic-Abliterated-Uncensored at Q4_K_S
- [Best LLM for RTX 3090](https://modelfit-eight.vercel.app/best-llm-for-rtx-3090/index.md): top pick Qwen3.8-27B-Heretic-Abliterated-Uncensored at Q4_K_S
- [Best LLM for RTX 5080](https://modelfit-eight.vercel.app/best-llm-for-rtx-5080/index.md): top pick Qwen3.8-27B-Heretic-Abliterated-Uncensored at Q2_K
- [Best LLM for RTX 4080](https://modelfit-eight.vercel.app/best-llm-for-rtx-4080/index.md): top pick Qwen3.8-27B-Heretic-Abliterated-Uncensored at Q2_K
- [Best LLM for RTX 4070 Ti Super](https://modelfit-eight.vercel.app/best-llm-for-rtx-4070-ti-super/index.md): top pick Qwen3.8-27B-Heretic-Abliterated-Uncensored at Q2_K
- [Best LLM for RTX 5070 Ti](https://modelfit-eight.vercel.app/best-llm-for-rtx-5070-ti/index.md): top pick Qwen3.8-27B-Heretic-Abliterated-Uncensored at Q2_K
- [Best LLM for RTX 4070 Super](https://modelfit-eight.vercel.app/best-llm-for-rtx-4070-super/index.md): top pick Qwen3.8-27B at UD-IQ2_XXS
- [Best LLM for RTX 4070](https://modelfit-eight.vercel.app/best-llm-for-rtx-4070/index.md): top pick Qwen3.8-27B at UD-IQ2_XXS
- [Best LLM for RTX 4060 Ti 16GB](https://modelfit-eight.vercel.app/best-llm-for-rtx-4060-ti-16gb/index.md): top pick Qwen3.8-27B-Heretic-Abliterated-Uncensored at Q2_K
- [Best LLM for RTX 3080 12GB](https://modelfit-eight.vercel.app/best-llm-for-rtx-3080/index.md): top pick Qwen3.8-27B at UD-IQ2_XXS
- [Best LLM for RTX 3060 12GB](https://modelfit-eight.vercel.app/best-llm-for-rtx-3060/index.md): top pick Qwen3.8-27B at UD-IQ2_XXS
- [Best LLM for RTX 4060 8GB](https://modelfit-eight.vercel.app/best-llm-for-rtx-4060/index.md): top pick Ternary-Bonsai-2-27B at BF16
- [Best LLM for RTX 3070 8GB](https://modelfit-eight.vercel.app/best-llm-for-rtx-3070/index.md): top pick Ternary-Bonsai-2-27B at BF16
- [Best LLM for MacBook Pro M4 (24GB)](https://modelfit-eight.vercel.app/best-llm-for-macbook-pro-m4-24gb/index.md): top pick Qwen3.8-27B-Heretic-Abliterated-Uncensored at Q3_K_S
- [Best LLM for MacBook Air M3/M4 (16GB)](https://modelfit-eight.vercel.app/best-llm-for-macbook-air-16gb/index.md): top pick Qwen3.8-27B at UD-IQ2_XXS
- [Best LLM for MacBook Air M2 (16GB)](https://modelfit-eight.vercel.app/best-llm-for-macbook-air-m2/index.md): top pick Qwen3.8-27B at UD-IQ2_XXS
- [Best LLM for MacBook Pro M4 Max (64GB)](https://modelfit-eight.vercel.app/best-llm-for-macbook-pro-m4-max/index.md): top pick Qwen3.8-27B-Heretic-Abliterated-Uncensored at IQ2_XXS
- [Best LLM for Mac Studio M3 Ultra (512GB)](https://modelfit-eight.vercel.app/best-llm-for-mac-studio-m3-ultra/index.md): top pick Kimi-K2.5 at UD-IQ1_S
- [Best LLM for 32GB system RAM (CPU inference)](https://modelfit-eight.vercel.app/best-llm-for-system-ram-32gb/index.md): top pick Qwen3.8-27B-Heretic-Abliterated-Uncensored at Q6_K
- [Best LLM for 64GB system RAM (CPU inference)](https://modelfit-eight.vercel.app/best-llm-for-system-ram-64gb/index.md): top pick Qwen3.8-27B-Heretic-Abliterated-Uncensored at IQ3_XXS
- [Best LLM for 128GB system RAM (CPU inference)](https://modelfit-eight.vercel.app/best-llm-for-system-ram-128gb/index.md): top pick GLM-5.3-Flash at UD-IQ1_S
- [What LLM can I run?](https://modelfit-eight.vercel.app/what-llm-can-i-run/): interactive VRAM/RAM picker over all catalogued models
- [Kimi-K3 vs deepseek-v4](https://modelfit-eight.vercel.app/compare-kimi-k3-vs-deepseek-v4/index.md): 561.1 GB vs 118.6 GB
- [deepseek-v4 vs gemma-4-E4B-it](https://modelfit-eight.vercel.app/compare-deepseek-v4-vs-gemma-4/index.md): 118.6 GB vs 11.1 GB
- [deepseek-v4 vs gemma-4-12b-it](https://modelfit-eight.vercel.app/compare-deepseek-v4-vs-gemma-4-12b/index.md): 118.6 GB vs 6.6 GB
- [Kimi-K3 vs Qwen3.8-27B](https://modelfit-eight.vercel.app/compare-kimi-k3-vs-qwen3-27b/index.md): 561.1 GB vs 13.0 GB
- [Qwen3.8-27B vs Qwen3-Coder-30B-A3B-Instruct](https://modelfit-eight.vercel.app/compare-qwen3-27b-vs-qwen3-coder-30b/index.md): 13.0 GB vs 13.1 GB
- [GLM-5.3 vs Qwen3.8-27B](https://modelfit-eight.vercel.app/compare-glm-5-3-vs-qwen3-27b/index.md): 261.6 GB vs 13.0 GB