Skip to content
AI Atlas
Hardware Fit tool

What fits my machine?

Pick a device memory size, a quantization and a context window: the atlas estimates each model's memory need from its parameter count. Every figure here is an estimate, never a measurement.

The custom field wins over the preset when both are filled. Estimates reserve 2 GB for the OS and framework.

EstimatedAll figures are estimates, not measurements.80 GB · 8-bit · 32k tokens540 of 658 models with a known parameter count fit

  • Estimated, not measured: weights = parameters × bytes/param × 1.15 runtime overhead.
  • bytes/param: 4bit = 0.5, 8bit = 1.0, fp16 = 2.0 (uniform quantization, no per-layer exceptions).
  • KV cache approximated at 0.5 GB per 8 192 tokens of context, independent of architecture (GQA/MLA models need less).
  • A model 'fits' when the estimate is at most the device memory minus 2 GB reserved for the OS and framework.
  • Mixture-of-experts models are estimated on total parameters (all experts must be resident); active parameters are ignored.
  • Device memory uses the largest configuration when several are listed (e.g. Apple silicon tiers).

Show only fitting modelsMethod →

Estimated hardware fit
ModelParamsQuantEst. memoryHeadroomFitsCompare
nvidia/Qwen3.5-122B-A10B-NVFP4Open weightsNVIDIAestimated: 64.6B params × 1.0 B × 1.15 + KV cache for 32768 tokens64.6B8bit76.3 GB+1.7 GB✓ fits
amd/gpt-oss-120b-w-mxfp4-a-fp8Open weightsAMDestimated: 59.5B params × 1.0 B × 1.15 + KV cache for 32768 tokens59.5B8bit70.4 GB+7.6 GB✓ fits
ai21labs/AI21-Jamba2-Mini-FP8Open weightsAI21 Labsestimated: 51.6B params × 1.0 B × 1.15 + KV cache for 32768 tokens51.6B8bit61.3 GB+16.7 GB✓ fits
ai21labs/AI21-Jamba-Mini-1.7-FP8RestrictedAI21 Labsestimated: 51.6B params × 1.0 B × 1.15 + KV cache for 32768 tokens51.6B8bit61.3 GB+16.7 GB✓ fits
Jamba-v0.1Open weightsAI21 Labsestimated: 51.6B params × 1.0 B × 1.15 + KV cache for 32768 tokens51.6B8bit61.3 GB+16.7 GB✓ fits
AI21-Jamba2-MiniOpen weightsAI21 Labsestimated: 51.6B params × 1.0 B × 1.15 + KV cache for 32768 tokens51.6B8bit61.3 GB+16.7 GB✓ fits
AI21-Jamba-Mini-1.7RestrictedAI21 Labsestimated: 51.6B params × 1.0 B × 1.15 + KV cache for 32768 tokens51.6B8bit61.3 GB+16.7 GB✓ fits
AI21-Jamba-Mini-1.5RestrictedAI21 Labsestimated: 51.6B params × 1.0 B × 1.15 + KV cache for 32768 tokens51.6B8bit61.3 GB+16.7 GB✓ fits
AI21-Jamba-Mini-1.6RestrictedAI21 Labsestimated: 51.6B params × 1.0 B × 1.15 + KV cache for 32768 tokens51.6B8bit61.3 GB+16.7 GB✓ fits
Kimi-Linear-48B-A3B-BaseOpen weightsMoonshot AIestimated: 49.1B params × 1.0 B × 1.15 + KV cache for 32768 tokens49.1B8bit58.5 GB+19.5 GB✓ fits
Kimi-Linear-48B-A3B-InstructOpen weightsMoonshot AIestimated: 49.1B params × 1.0 B × 1.15 + KV cache for 32768 tokens49.1B8bit58.5 GB+19.5 GB✓ fits
function-calling-v1Open weightsFireworks AIestimated: 46.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens46.7B8bit55.7 GB+22.3 GB✓ fits
firefunction-v1Open weightsFireworks AIestimated: 46.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens46.7B8bit55.7 GB+22.3 GB✓ fits
Nous-Hermes-2-Mixtral-8x7B-DPOOpen weightsNous Researchestimated: 46.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens46.7B8bit55.7 GB+22.3 GB✓ fits
Mixtral-8x7B-Instruct-v0.1Open weightsMistral AIestimated: 46.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens46.7B8bit55.7 GB+22.3 GB✓ fits
DFN2B-CLIP-ViT-L-14-39BOpen weightsAppleestimated: 39.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens39B8bit46.9 GB+31.1 GB✓ fits
Seed-OSS-36B-BaseOpen weightsByteDanceestimated: 36.2B params × 1.0 B × 1.15 + KV cache for 32768 tokens36.1B8bit43.6 GB+34.4 GB✓ fits
Seed-OSS-36B-InstructOpen weightsByteDanceestimated: 36.2B params × 1.0 B × 1.15 + KV cache for 32768 tokens36.1B8bit43.6 GB+34.4 GB✓ fits
NousResearch/Hermes-4.3-36B-GGUFOpen weightsNous Researchestimated: 36.2B params × 1.0 B × 1.15 + KV cache for 32768 tokens36.1B8bit43.6 GB+34.4 GB✓ fits
Qwen/Qwen3.6-35B-A3B-FP8Open weightsQwenestimated: 36.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens36B8bit43.4 GB+34.6 GB✓ fits
Qwen3.6 35B A3BOpen weightsQwenestimated: 36.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens36B8bit43.3 GB+34.7 GB✓ fits
unsloth/Qwen3.6-35B-A3B-MTP-GGUFOpen weightsUnslothestimated: 35.5B params × 1.0 B × 1.15 + KV cache for 32768 tokens35.5B8bit42.8 GB+35.2 GB✓ fits
bartowski/endless-frontier_BigBang-v1-GGUFOpen weightsbartowskiestimated: 35.5B params × 1.0 B × 1.15 + KV cache for 32768 tokens35.5B8bit42.8 GB+35.2 GB✓ fits
cerebras/Kimi-Linear-REAP-35B-A3B-InstructOpen weightsCerebras Systemsestimated: 35.1B params × 1.0 B × 1.15 + KV cache for 32768 tokens35.1B8bit42.4 GB+35.6 GB✓ fits
perplexity-ai/pplx-computer-qwen-3-6-35b-a3b-nvfp4-20260709Open weightsPerplexity AIestimated: 35.1B params × 1.0 B × 1.15 + KV cache for 32768 tokens35.1B8bit42.4 GB+35.6 GB✓ fits
bartowski/Qwen_Qwen3.6-35B-A3B-GGUFOpen weightsbartowskiestimated: 35.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens35B8bit42.3 GB+35.8 GB✓ fits
bartowski/Qwen_Qwen3.5-35B-A3B-GGUFOpen weightsbartowskiestimated: 35.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens35B8bit42.3 GB+35.8 GB✓ fits
c4ai-command-r-v01RestrictedCohereestimated: 35.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens35B8bit42.2 GB+35.8 GB✓ fits
bartowski/Kwaipilot_KAT-Coder-V2.5-Dev-GGUFOpen weightsbartowskiestimated: 34.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens34.7B8bit41.9 GB+36.1 GB✓ fits
bartowski/thomsonreuters_Thomson-1.0-Small-GGUFOpen weightsbartowskiestimated: 34.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens34.7B8bit41.9 GB+36.1 GB✓ fits
unsloth/Qwen3.6-35B-A3B-GGUFOpen weightsUnslothestimated: 34.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens34.7B8bit41.9 GB+36.1 GB✓ fits
bartowski/XYZAILab_XYZ-Aquila-mini-GGUFOpen weightsbartowskiestimated: 34.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens34.7B8bit41.9 GB+36.1 GB✓ fits
perplexity-ai/pplx-computer-qwen-3-6-35b-a3b-mlx-20260709Open weightsPerplexity AIestimated: 34.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens34.7B8bit41.9 GB+36.1 GB✓ fits
Nous-Hermes-2-Yi-34BOpen weightsNous Researchestimated: 34.4B params × 1.0 B × 1.15 + KV cache for 32768 tokens34.4B8bit41.5 GB+36.5 GB✓ fits
GKA-primed-HQwen3-32B-ReasonerOpen weightsAmazon Web Servicesestimated: 34.1B params × 1.0 B × 1.15 + KV cache for 32768 tokens34.1B8bit41.3 GB+36.7 GB✓ fits
MiniMax-H3Open weightsMiniMaxestimated: 33.1B params × 1.0 B × 1.15 + KV cache for 32768 tokens33.1B8bit40.1 GB+37.9 GB✓ fits
nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-FP8Open weightsNVIDIAestimated: 33.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens33B8bit40.0 GB+38.0 GB✓ fits
SynLogic-32BOpen weightsMiniMaxestimated: 32.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens32.8B8bit39.7 GB+38.3 GB✓ fits
DeepSeek-R1-Distill-Qwen-32BOpen weightsDeepSeekestimated: 32.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens32.8B8bit39.7 GB+38.3 GB✓ fits
SynLogic-Mix-3-32BOpen weightsMiniMaxestimated: 32.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens32.8B8bit39.7 GB+38.3 GB✓ fits
Qwen3 32BOpen weightsQwenestimated: 32.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens32.8B8bit39.7 GB+38.3 GB✓ fits
c4ai-command-r-08-2024RestrictedCohereestimated: 32.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens32.3B8bit39.1 GB+38.9 GB✓ fits
aya-expanse-32bRestrictedCohereestimated: 32.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens32.3B8bit39.1 GB+38.9 GB✓ fits
OLMo-2-0325-32B-InstructOpen weightsAllen Institute for AIestimated: 32.2B params × 1.0 B × 1.15 + KV cache for 32768 tokens32.2B8bit39.1 GB+38.9 GB✓ fits
FLUX.2-devRestrictedBlack Forest Labsestimated: 32.2B params × 1.0 B × 1.15 + KV cache for 32768 tokens32.2B8bit39.1 GB+38.9 GB✓ fits
Nemotron 3 Nano 30B A3BOpen weightsNVIDIAestimated: 31.6B params × 1.0 B × 1.15 + KV cache for 32768 tokens31.6B8bit38.3 GB+39.7 GB✓ fits
mlx-community/gemma-4-31b-it-4bitOpen weightsMLX Communityestimated: 31.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens31.3B8bit38.0 GB+40.0 GB✓ fits
Gemma 4 31BOpen weightsGoogleestimated: 31.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens31.3B8bit38.0 GB+40.0 GB✓ fits
GLM 4.7 FlashOpen weightsZ.ai (Zhipu AI)estimated: 31.2B params × 1.0 B × 1.15 + KV cache for 32768 tokens31.2B8bit37.9 GB+40.1 GB✓ fits
unsloth/gemma-4-31B-it-qat-GGUFOpen weightsUnslothestimated: 30.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens30.7B8bit37.3 GB+40.7 GB✓ fits
mlx-community/Qwen3-30B-A3B-Instruct-2507-4bitOpen weightsMLX Communityestimated: 30.5B params × 1.0 B × 1.15 + KV cache for 32768 tokens30.5B8bit37.1 GB+40.9 GB✓ fits
browsesafeOpen weightsPerplexity AIestimated: 30.5B params × 1.0 B × 1.15 + KV cache for 32768 tokens30.5B8bit37.1 GB+40.9 GB✓ fits
unsloth/Qwen3-30B-A3B-Thinking-2507-GGUFOpen weightsUnslothestimated: 30.5B params × 1.0 B × 1.15 + KV cache for 32768 tokens30.5B8bit37.1 GB+40.9 GB✓ fits
unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUFOpen weightsUnslothestimated: 30.5B params × 1.0 B × 1.15 + KV cache for 32768 tokens30.5B8bit37.1 GB+40.9 GB✓ fits
North Mini Code (free)Open weightsCohereestimated: 30.5B params × 1.0 B × 1.15 + KV cache for 32768 tokens30.5B8bit37.1 GB+40.9 GB✓ fits
Hy-MT2-30B-A3BOpen weightsTencentestimated: 30.1B params × 1.0 B × 1.15 + KV cache for 32768 tokens30.1B8bit36.6 GB+41.4 GB✓ fits
north-mini-code-1-0Cohereestimated: 30.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens30B8bit36.5 GB+41.5 GB✓ fits
ibm-granite/granite-4.2-30b-GGUFOpen weightsIBMestimated: 30.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens30B8bit36.5 GB+41.5 GB✓ fits
ERNIE-4.5-VL-28B-A3B-ThinkingOpen weightsBaiduestimated: 29.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens29.7B8bit36.1 GB+41.9 GB✓ fits
ERNIE-4.5-VL-28B-A3B-PTOpen weightsBaiduestimated: 29.4B params × 1.0 B × 1.15 + KV cache for 32768 tokens29.4B8bit35.8 GB+42.2 GB✓ fits
granite-4.1-30bOpen weightsIBMestimated: 28.9B params × 1.0 B × 1.15 + KV cache for 32768 tokens28.9B8bit35.2 GB+42.8 GB✓ fits
Qwen/Qwen3.6-27B-FP8Open weightsQwenestimated: 27.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens27.8B8bit34.0 GB+44.0 GB✓ fits
Qwen/Qwen3.8-27B-FP8Open weightsQwenestimated: 27.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens27.8B8bit34.0 GB+44.0 GB✓ fits
Qwen3.8 27BOpen weightsQwenestimated: 27.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens27.8B8bit34.0 GB+44.0 GB✓ fits
Qwen3.6 27BOpen weightsQwenestimated: 27.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens27.8B8bit34.0 GB+44.0 GB✓ fits
mlx-community/Qwen3.8-27B-8bitOpen weightsMLX Communityestimated: 27.4B params × 1.0 B × 1.15 + KV cache for 32768 tokens27.4B8bit33.5 GB+44.5 GB✓ fits
mlx-community/Qwen3.8-27B-4bitOpen weightsMLX Communityestimated: 27.4B params × 1.0 B × 1.15 + KV cache for 32768 tokens27.4B8bit33.5 GB+44.5 GB✓ fits
unsloth/Qwen3.8-27B-GGUFOpen weightsUnslothestimated: 27.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens27.3B8bit33.4 GB+44.6 GB✓ fits
perplexity-ai/pplx-computer-qwen-3-8-27b-dflash2-gguf-20260826Open weightsPerplexity AIestimated: 27.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens27.3B8bit33.4 GB+44.6 GB✓ fits
unsloth/Qwen3.6-27B-MTP-GGUFOpen weightsUnslothestimated: 27.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens27.3B8bit33.4 GB+44.6 GB✓ fits
bartowski/Fara1.5-27B-GGUFOpen weightsbartowskiestimated: 27.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens27B8bit33.0 GB+45.0 GB✓ fits
bartowski/Qwen3.8-27B-GGUFOpen weightsbartowskiestimated: 27.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens27B8bit33.0 GB+45.0 GB✓ fits
unsloth/Qwen3.6-27B-GGUFOpen weightsUnslothestimated: 26.9B params × 1.0 B × 1.15 + KV cache for 32768 tokens26.9B8bit32.9 GB+45.1 GB✓ fits
Gemma 4 26B A4BOpen weightsGoogleestimated: 25.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens25.8B8bit31.7 GB+46.3 GB✓ fits
unsloth/gemma-4-26B-A4B-it-qat-GGUFOpen weightsUnslothestimated: 25.2B params × 1.0 B × 1.15 + KV cache for 32768 tokens25.2B8bit31.0 GB+47.0 GB✓ fits
unsloth/gemma-4-26B-A4B-it-GGUFOpen weightsUnslothestimated: 25.2B params × 1.0 B × 1.15 + KV cache for 32768 tokens25.2B8bit31.0 GB+47.0 GB✓ fits
cerebras/Qwen3-Coder-REAP-25B-A3BOpen weightsCerebras Systemsestimated: 24.9B params × 1.0 B × 1.15 + KV cache for 32768 tokens24.9B8bit30.6 GB+47.4 GB✓ fits
unsloth/Qwen3.6-35B-A3B-NVFP4Open weightsUnslothestimated: 24.6B params × 1.0 B × 1.15 + KV cache for 32768 tokens24.6B8bit30.3 GB+47.7 GB✓ fits
Voxtral Small 24B 2507Open weightsMistral AIestimated: 24.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens24.3B8bit29.9 GB+48.1 GB✓ fits
Devstral-Small-2-24B-Instruct-2512Open weightsMistral AIestimated: 24.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens24B8bit29.6 GB+48.4 GB✓ fits
Mistral Small 3.1 24BOpen weightsMistral AIestimated: 24.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens24B8bit29.6 GB+48.4 GB✓ fits
Mistral Small 3.2 24BOpen weightsMistral AIestimated: 24.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens24B8bit29.6 GB+48.4 GB✓ fits
mlx-community/Devstral-Small-2-24B-Instruct-2512-4bitOpen weightsMLX Communityestimated: 24.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens24B8bit29.6 GB+48.4 GB✓ fits
cerebras/GLM-4.7-Flash-REAP-23B-A3BOpen weightsCerebras Systemsestimated: 23.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens23B8bit28.4 GB+49.5 GB✓ fits
ERNIE-4.5-21B-A3B-PTOpen weightsBaiduestimated: 21.9B params × 1.0 B × 1.15 + KV cache for 32768 tokens21.9B8bit27.2 GB+50.8 GB✓ fits
ERNIE-4.5-21B-A3B-Base-PTOpen weightsBaiduestimated: 21.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens21.8B8bit27.1 GB+50.9 GB✓ fits
ERNIE-4.5-21B-A3B-ThinkingOpen weightsBaiduestimated: 21.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens21.8B8bit27.1 GB+50.9 GB✓ fits
gpt-oss-safeguard-20bOpen weightsOpenAIestimated: 21.5B params × 1.0 B × 1.15 + KV cache for 32768 tokens21.5B8bit26.7 GB+51.3 GB✓ fits
unsloth/Qwen3.6-27B-NVFP4Open weightsUnslothestimated: 21.2B params × 1.0 B × 1.15 + KV cache for 32768 tokens21.2B8bit26.4 GB+51.6 GB✓ fits
gpt-oss-20bOpen weightsOpenAIestimated: 20.9B params × 1.0 B × 1.15 + KV cache for 32768 tokens20.9B8bit26.1 GB+52.0 GB✓ fits
mlx-community/gpt-oss-20b-MXFP4-Q8Open weightsMLX Communityestimated: 20.9B params × 1.0 B × 1.15 + KV cache for 32768 tokens20.9B8bit26.1 GB+52.0 GB✓ fits
nvidia/Gemma-4-31B-IT-NVFP4Open weightsNVIDIAestimated: 20.9B params × 1.0 B × 1.15 + KV cache for 32768 tokens20.9B8bit26.0 GB+52.0 GB✓ fits
gpt-neox-20bOpen weightsEleutherAIestimated: 20.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens20.7B8bit25.9 GB+52.1 GB✓ fits
unsloth/MiniMax-H3-GGUFOpen weightsUnslothestimated: 20.1B params × 1.0 B × 1.15 + KV cache for 32768 tokens20.1B8bit25.1 GB+52.9 GB✓ fits
internlm2-20bOpen weightsInternLM (Shanghai AI Laboratory)estimated: 20.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens20B8bit25.0 GB+53.0 GB✓ fits
internlm/internlm2_5-20b-chat-ggufOpen weightsInternLM (Shanghai AI Laboratory)estimated: 20.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens20B8bit25.0 GB+53.0 GB✓ fits
internlm2-base-20bOpen weightsInternLM (Shanghai AI Laboratory)estimated: 20.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens20B8bit25.0 GB+53.0 GB✓ fits
unsloth/Qwen3.8-27B-NVFP4Open weightsUnslothestimated: 19.9B params × 1.0 B × 1.15 + KV cache for 32768 tokens19.9B8bit24.9 GB+53.1 GB✓ fits
internlm2-chat-20bOpen weightsInternLM (Shanghai AI Laboratory)estimated: 19.9B params × 1.0 B × 1.15 + KV cache for 32768 tokens19.9B8bit24.8 GB+53.2 GB✓ fits
Intel/Qwen3.5-122B-A10B-int4-AutoRoundOpen weightsIntelestimated: 19.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens19.7B8bit24.6 GB+53.4 GB✓ fits
pplx-qwen-3-8-27b-dflash2-20260819Open weightsPerplexity AIestimated: 18.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens18.8B8bit23.6 GB+54.4 GB✓ fits
pplx-computer-qwen-3-8-27b-dflash2-20260824Open weightsPerplexity AIestimated: 18.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens18.8B8bit23.6 GB+54.4 GB✓ fits
nvidia/Qwen3.6-35B-A3B-NVFP4Open weightsNVIDIAestimated: 18.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens18.7B8bit23.5 GB+54.5 GB✓ fits
nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-NVFP4Open weightsNVIDIAestimated: 18.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens18.3B8bit23.1 GB+54.9 GB✓ fits
nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4Open weightsNVIDIAestimated: 17.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens17.8B8bit22.5 GB+55.5 GB✓ fits
Kimi-VL-A3B-ThinkingOpen weightsMoonshot AIestimated: 16.4B params × 1.0 B × 1.15 + KV cache for 32768 tokens16.4B8bit20.9 GB+57.1 GB✓ fits
Kimi-VL-A3B-Thinking-2506Open weightsMoonshot AIestimated: 16.4B params × 1.0 B × 1.15 + KV cache for 32768 tokens16.4B8bit20.9 GB+57.1 GB✓ fits
Kimi-VL-A3B-InstructOpen weightsMoonshot AIestimated: 16.4B params × 1.0 B × 1.15 + KV cache for 32768 tokens16.4B8bit20.9 GB+57.1 GB✓ fits
Moonlight-16B-A3B-InstructOpen weightsMoonshot AIestimated: 16.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens16B8bit20.4 GB+57.6 GB✓ fits
Moonlight-16B-A3BOpen weightsMoonshot AIestimated: 16.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens16B8bit20.4 GB+57.6 GB✓ fits
bartowski/DeepSeek-Coder-V2-Lite-Instruct-GGUFOpen weightsbartowskiestimated: 15.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens15.7B8bit20.1 GB+57.9 GB✓ fits
DeepSeek-V2-LiteOpen weightsDeepSeekestimated: 15.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens15.7B8bit20.1 GB+57.9 GB✓ fits
DeepSeek-Coder-V2-Lite-InstructOpen weightsDeepSeekestimated: 15.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens15.7B8bit20.1 GB+57.9 GB✓ fits
amd/Qwen3.8-27B-Quark-AWQ-MXFP4Open weightsAMDestimated: 15.6B params × 1.0 B × 1.15 + KV cache for 32768 tokens15.6B8bit19.9 GB+58.0 GB✓ fits
BAGEL-7B-MoTOpen weightsByteDanceestimated: 14.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens14.7B8bit18.9 GB+59.1 GB✓ fits
Phi 4Open weightsMicrosoftestimated: 14.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens14.7B8bit18.9 GB+59.1 GB✓ fits
nvidia/Gemma-4-26B-A4B-NVFP4Open weightsNVIDIAestimated: 14.4B params × 1.0 B × 1.15 + KV cache for 32768 tokens14.4B8bit18.5 GB+59.5 GB✓ fits
Ministral 3 14B 2512Open weightsMistral AIestimated: 13.9B params × 1.0 B × 1.15 + KV cache for 32768 tokens13.9B8bit18.0 GB+60.0 GB✓ fits
Ministral-3-14B-Reasoning-2512Open weightsMistral AIestimated: 13.9B params × 1.0 B × 1.15 + KV cache for 32768 tokens13.9B8bit18.0 GB+60.0 GB✓ fits
FireLLaVA-13bRestrictedFireworks AIestimated: 13.4B params × 1.0 B × 1.15 + KV cache for 32768 tokens13.3B8bit17.4 GB+60.6 GB✓ fits
gemma-4-12B-it-qat-w4a16-ctOpen weightsGoogleestimated: 13.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens13.3B8bit17.3 GB+60.7 GB✓ fits
Llama-2-13b-chat-hfRestrictedMeta AIestimated: 13.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens13B8bit17.0 GB+61.0 GB✓ fits
aya-101Open weightsCohereestimated: 12.9B params × 1.0 B × 1.15 + KV cache for 32768 tokens12.9B8bit16.9 GB+61.1 GB✓ fits
Mistral NemoOpen weightsMistral AIestimated: 12.2B params × 1.0 B × 1.15 + KV cache for 32768 tokens12.3B8bit16.1 GB+61.9 GB✓ fits
Mistral-Nemo-Base-2407Open weightsMistral AIestimated: 12.2B params × 1.0 B × 1.15 + KV cache for 32768 tokens12.3B8bit16.1 GB+61.9 GB✓ fits
bartowski/kai-os_Grug-12B-GGUFOpen weightsbartowskiestimated: 12.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens12B8bit15.8 GB+62.2 GB✓ fits
mlx-community/gemma-3-12b-it-4bitOpen weightsMLX Communityestimated: 12.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens12B8bit15.8 GB+62.2 GB✓ fits
gemma-4-12B-itOpen weightsGoogleestimated: 12.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens12B8bit15.8 GB+62.3 GB✓ fits
unsloth/gemma-4-12B-it-qat-GGUFOpen weightsUnslothestimated: 11.9B params × 1.0 B × 1.15 + KV cache for 32768 tokens11.9B8bit15.7 GB+62.3 GB✓ fits
FLUX.1-Fill-devRestrictedBlack Forest Labsestimated: 11.9B params × 1.0 B × 1.15 + KV cache for 32768 tokens11.9B8bit15.7 GB+62.3 GB✓ fits
FLUX.1-Krea-devRestrictedBlack Forest Labsestimated: 11.9B params × 1.0 B × 1.15 + KV cache for 32768 tokens11.9B8bit15.7 GB+62.3 GB✓ fits
FLUX.1-Kontext-devRestrictedBlack Forest Labsestimated: 11.9B params × 1.0 B × 1.15 + KV cache for 32768 tokens11.9B8bit15.7 GB+62.3 GB✓ fits
FLUX.1-devRestrictedBlack Forest Labsestimated: 11.9B params × 1.0 B × 1.15 + KV cache for 32768 tokens11.9B8bit15.7 GB+62.3 GB✓ fits
FLUX.1-schnellRestrictedBlack Forest Labsestimated: 11.9B params × 1.0 B × 1.15 + KV cache for 32768 tokens11.9B8bit15.7 GB+62.3 GB✓ fits
Intel/Qwen3-Coder-Next-int4-AutoRoundOpen weightsIntelestimated: 11.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens11.8B8bit15.6 GB+62.4 GB✓ fits
KaLM-Embedding-Gemma3-12B-2511Open weightsTencentestimated: 11.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens11.8B8bit15.5 GB+62.5 GB✓ fits
Nous-Hermes-2-SOLAR-10.7BOpen weightsNous Researchestimated: 10.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens10.7B8bit14.3 GB+63.7 GB✓ fits
Llama-3.2-11B-Vision-InstructRestrictedMeta AIestimated: 10.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens10.7B8bit14.3 GB+63.7 GB✓ fits
GLM-4.1V-9B-ThinkingOpen weightsZ.ai (Zhipu AI)estimated: 10.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens10.3B8bit13.8 GB+64.2 GB✓ fits
GLM-4.6V-FlashOpen weightsZ.ai (Zhipu AI)estimated: 10.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens10.3B8bit13.8 GB+64.2 GB✓ fits
Kimi-Audio-7BOpen weightsMoonshot AIestimated: 9.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens9.77B8bit13.2 GB+64.8 GB✓ fits
Kimi-Audio-7B-InstructOpen weightsMoonshot AIestimated: 9.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens9.77B8bit13.2 GB+64.8 GB✓ fits
Qwen3.5-9BOpen weightsQwenestimated: 9.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens9.65B8bit13.1 GB+64.9 GB✓ fits
glm-4-9b-chatOpen weightsZ.ai (Zhipu AI)estimated: 9.4B params × 1.0 B × 1.15 + KV cache for 32768 tokens9.4B8bit12.8 GB+65.2 GB✓ fits
academic-ds-9BOpen weightsByteDanceestimated: 9.4B params × 1.0 B × 1.15 + KV cache for 32768 tokens9.37B8bit12.8 GB+65.2 GB✓ fits
FLUX.2-klein-base-9BRestrictedBlack Forest Labsestimated: 9.1B params × 1.0 B × 1.15 + KV cache for 32768 tokens9.08B8bit12.4 GB+65.6 GB✓ fits
FLUX.2-klein-9b-kvRestrictedBlack Forest Labsestimated: 9.1B params × 1.0 B × 1.15 + KV cache for 32768 tokens9.08B8bit12.4 GB+65.6 GB✓ fits
FLUX.2-klein-9BRestrictedBlack Forest Labsestimated: 9.1B params × 1.0 B × 1.15 + KV cache for 32768 tokens9.08B8bit12.4 GB+65.6 GB✓ fits
bartowski/tencent_UI-Mate-9B-GGUFOpen weightsbartowskiestimated: 9.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens9B8bit12.3 GB+65.7 GB✓ fits
black-forest-labs/FLUX.2-klein-9b-kv-fp8Open weightsBlack Forest Labsestimated: 9.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens9B8bit12.3 GB+65.7 GB✓ fits
black-forest-labs/FLUX.2-klein-base-9b-fp8RestrictedBlack Forest Labsestimated: 9.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens9B8bit12.3 GB+65.7 GB✓ fits
bartowski/Fara1.5-9B-GGUFOpen weightsbartowskiestimated: 9.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens9B8bit12.3 GB+65.7 GB✓ fits
black-forest-labs/FLUX.2-klein-9b-fp8RestrictedBlack Forest Labsestimated: 9.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens9B8bit12.3 GB+65.7 GB✓ fits
unsloth/Qwen3.5-9B-GGUFOpen weightsUnslothestimated: 9.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.95B8bit12.3 GB+65.7 GB✓ fits
Qianfan-VL-8BOpen weightsBaiduestimated: 8.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.81B8bit12.1 GB+65.9 GB✓ fits
internlm3-8b-instructOpen weightsInternLM (Shanghai AI Laboratory)estimated: 8.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.8B8bit12.1 GB+65.9 GB✓ fits
granite-4.1-8bOpen weightsIBMestimated: 8.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.79B8bit12.1 GB+65.9 GB✓ fits
amd/Qwen3-VL-8B-Instruct-w8a8-llmcompressorOpen weightsAMDestimated: 8.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.77B8bit12.1 GB+65.9 GB✓ fits
Qwen3 VL 8B InstructOpen weightsQwenestimated: 8.8B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.77B8bit12.1 GB+65.9 GB✓ fits
VibeVoice-ASROpen weightsMicrosoftestimated: 8.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.67B8bit12.0 GB+66.0 GB✓ fits
Molmo2-8BOpen weightsAllen Institute for AIestimated: 8.7B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.66B8bit12.0 GB+66.0 GB✓ fits
aya-vision-8bRestrictedCohereestimated: 8.6B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.63B8bit11.9 GB+66.1 GB✓ fits
Intern-S1-miniOpen weightsInternLM (Shanghai AI Laboratory)estimated: 8.5B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.54B8bit11.8 GB+66.2 GB✓ fits
GKA-primed-HQwen3-8B-ReasonerOpen weightsAmazon Web Servicesestimated: 8.5B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.5B8bit11.8 GB+66.2 GB✓ fits
GDN-primed-HQwen3-8B-InstructOpen weightsAmazon Web Servicesestimated: 8.5B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.5B8bit11.8 GB+66.2 GB✓ fits
LFM2.5-8B-A1BOpen weightsLiquid AIestimated: 8.5B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.47B8bit11.7 GB+66.3 GB✓ fits
allenai/olmOCR-2-7B-1025-FP8Open weightsAllen Institute for AIestimated: 8.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.29B8bit11.5 GB+66.5 GB✓ fits
UI-TARS 7BOpen weightsByteDanceestimated: 8.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.29B8bit11.5 GB+66.5 GB✓ fits
olmOCR-2-7B-1025Open weightsAllen Institute for AIestimated: 8.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.29B8bit11.5 GB+66.5 GB✓ fits
Qwen2.5-VL-7B-InstructOpen weightsQwenestimated: 8.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.29B8bit11.5 GB+66.5 GB✓ fits
UI-TARS-7B-DPOOpen weightsByteDanceestimated: 8.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.29B8bit11.5 GB+66.5 GB✓ fits
UI-TARS-7B-SFTOpen weightsByteDanceestimated: 8.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.29B8bit11.5 GB+66.5 GB✓ fits
Stable-DiffCoder-8B-InstructOpen weightsByteDanceestimated: 8.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.25B8bit11.5 GB+66.5 GB✓ fits
Seed-Coder-8B-ReasoningOpen weightsByteDanceestimated: 8.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.25B8bit11.5 GB+66.5 GB✓ fits
Stable-DiffCoder-8B-BaseOpen weightsByteDanceestimated: 8.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.25B8bit11.5 GB+66.5 GB✓ fits
Seed-Coder-8B-BaseOpen weightsByteDanceestimated: 8.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.25B8bit11.5 GB+66.5 GB✓ fits
Seed-Coder-8B-InstructOpen weightsByteDanceestimated: 8.3B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.25B8bit11.5 GB+66.5 GB✓ fits
Qwen3 8BOpen weightsQwenestimated: 8.2B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.19B8bit11.4 GB+66.6 GB✓ fits
DeepSeek-R1-0528-Qwen3-8BOpen weightsDeepSeekestimated: 8.2B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.19B8bit11.4 GB+66.6 GB✓ fits
mlx-community/Qwen3-8B-4bitOpen weightsMLX Communityestimated: 8.2B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.19B8bit11.4 GB+66.6 GB✓ fits
granite-guardian-3.3-8bOpen weightsIBMestimated: 8.2B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.17B8bit11.4 GB+66.6 GB✓ fits
granite-3.0-8b-instructOpen weightsIBMestimated: 8.2B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.17B8bit11.4 GB+66.6 GB✓ fits
stable-diffusion-3.5-largeRestrictedStability AIestimated: 8.1B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.15B8bit11.4 GB+66.6 GB✓ fits
ERNIE-Image-TurboOpen weightsBaiduestimated: 8.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.03B8bit11.2 GB+66.8 GB✓ fits
ERNIE-ImageOpen weightsBaiduestimated: 8.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.03B8bit11.2 GB+66.8 GB✓ fits
NousResearch/Hermes-3-Llama-3.1-8B-GGUFOpen weightsNous Researchestimated: 8.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.03B8bit11.2 GB+66.8 GB✓ fits
Hermes-3-Llama-3.1-8BOpen weightsNous Researchestimated: 8.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.03B8bit11.2 GB+66.8 GB✓ fits
NousResearch/Meta-Llama-3-8BOpen weightsNous Researchestimated: 8.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.03B8bit11.2 GB+66.8 GB✓ fits
Llama-3.1-8B-InstructRestrictedMeta AIestimated: 8.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.03B8bit11.2 GB+66.8 GB✓ fits
Meta-Llama-3-8BRestrictedMeta AIestimated: 8.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.03B8bit11.2 GB+66.8 GB✓ fits
NousResearch/Meta-Llama-3.1-8B-InstructOpen weightsNous Researchestimated: 8.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.03B8bit11.2 GB+66.8 GB✓ fits
NousResearch/Meta-Llama-3-8B-InstructOpen weightsNous Researchestimated: 8.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.03B8bit11.2 GB+66.8 GB✓ fits
NousResearch/Meta-Llama-3.1-8BOpen weightsNous Researchestimated: 8.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.03B8bit11.2 GB+66.8 GB✓ fits
Llama 3.1 8BRestrictedMeta AIestimated: 8.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.03B8bit11.2 GB+66.8 GB✓ fits
Salesforce/Llama-xLAM-2-8b-fc-rOpen weightsSalesforceestimated: 8.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.03B8bit11.2 GB+66.8 GB✓ fits
Hermes-2-Theta-Llama-3-8BOpen weightsNous Researchestimated: 8.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.03B8bit11.2 GB+66.8 GB✓ fits
mlx-community/Llama-3.1-8B-Instruct-4bitOpen weightsMLX Communityestimated: 8.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.03B8bit11.2 GB+66.8 GB✓ fits
Meta-Llama-3-8B-InstructRestrictedMeta AIestimated: 8.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.03B8bit11.2 GB+66.8 GB✓ fits
Hy-MT2-7BOpen weightsTencentestimated: 8.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.03B8bit11.2 GB+66.8 GB✓ fits
aya-23-8BRestrictedCohereestimated: 8.0B params × 1.0 B × 1.15 + KV cache for 32768 tokens8.03B8bit11.2 GB+66.8 GB✓ fits

Headroom = device memory − 2 GB reserve − estimated need. Sorted by the API (fitting models first). Compare the models you shortlist with the + buttons. Devices with these memory sizes: hardware listing.