RunsOnMy ModelsGPUs & MacsAPI pricesVRAM calculatorNew

What AI models run on the A100 80GB?

The A100 80GB has 80 GB of VRAM (2,039 GB/s). About 76.0 GB is usable for a model. 220 of 285 open models we track fit at 8K context.

220models fit (8K ctx)
218still fit at 32K ctx
201Blargest that fits
2039 GB/smemory bandwidth

Models that fit, largest first

ModelParamsBest quant (8K)Needs32K contextMax tok/s (est.)
Step 3.7 Flash201BQ2_K74.5 GBno27details
Step 3.5 Flash199BQ2_K69.0 GBno29details
Qwen3.8 Flash Next180B (62B act.)Q2_K63.9 GBQ2_K88details
Ling 3.0 flash Sante127B (5.6B act.)Q3_K_M58.3 GBQ3_K_M754details
Ling 3.0 flash127B (5.6B act.)Q3_K_M56.0 GBQ3_K_M786details
Ornith 1.5 122B125B (13B act.)Q3_K_M58.0 GBQ3_K_M327details
gpt oss 120b117B (5.8B act.)F1661.9 GBF16634details
gpt oss safeguard 120b117B (5.8B act.)Q8_060.1 GBQ8_0654details
GLM 4.5 Air110B (15B act.)Q3_K_M55.4 GBQ3_K_M263details
Llama 4 Scout 17B 16E Instruct109B (18B act.)Q4_K_M63.2 GBQ4_K_M188details
Llama 4 Scout 17B 16E109B (18B act.)Q4_K_M64.1 GBQ3_K_M185details
GLM 4.5V108B (14B act.)Q4_K_M61.4 GBQ4_K_M242details
GLM 4.6V108B (14B act.)Q4_K_M67.8 GBQ3_K_M218details
Hunyuan A13B Instruct80B (13B act.)Q6_K63.0 GBQ6_K191details
Kimi Dev 72B73BQ6_K63.1 GBQ5_K_M32details
Llama 3.3 70B Instruct71BQ6_K57.0 GBQ6_K35details
DeepSeek R1 Distill Llama 70B71BQ6_K57.0 GBQ6_K35details
Llama 3.1 70B71BQ6_K57.1 GBQ6_K35details
Kimi Linear 48B A3B Instruct49B (3.0B act.)Q8_049.6 GBQ8_0639details
Nous Hermes 2 Mixtral 8x7B DPO47B (13B act.)Q8_047.7 GBQ8_0149details
falcon 40b42BQ8_042.9 GBQ8_046details
flint40B (1.3B act.)Q8_040.3 GBQ8_01534details
Maxwell V1 Flash37B (2.3B act.)Q8_037.6 GBQ8_0845details
Hermes 4.3 36B36BQ8_038.5 GBQ8_053details
Seed OSS 36B Instruct36BQ8_038.5 GBQ8_053details
Qwen3.6 35B A3B36B (3.0B act.)F1667.2 GBF16344details
Ornith 1.5 35B A3B36B (3.0B act.)F1667.2 GBF16344details
Qwen3.5 35B A3B36B (3.0B act.)F1665.6 GBF16352details
Axion 1.0 35B35B (3.9B act.)F1665.6 GBF16264details
Qwen AgentWorld 35B A3B35B (3.0B act.)F1665.6 GBF16340details
Yi 34B 200K34BF1666.5 GBQ8_030details
Yi 34B34BF1666.5 GBQ8_030details
Yi 34B Chat34BF1666.5 GBQ8_030details
Yi 1.5 34B Chat34BF1666.5 GBQ8_030details
Nous Hermes 2 Yi 34B34BF1666.5 GBQ8_030details
DeepSeek R1 Distill Qwen 32B33BF1663.7 GBQ8_031details
Qwen3 32B33BF1663.7 GBQ8_031details
GLM Z1 32B 041433BF1661.8 GBF1631details
GLM 4 32B 041433BF1661.8 GBF1631details
granite 4.0 h small32B (8.8B act.)F1660.7 GBF16116details
gemma 4 31B it31BF1660.1 GBF1633details
GLM 4.7 Flash31B (4.6B act.)F1656.9 GBF16231details
Qwen3 Coder 30B A3B Instruct31B (3.0B act.)F1658.3 GBF16340details
North Mini Code 1.030B (3.3B act.)F1658.2 GBF16308details
Hy Memory 30B A3B30B (3.0B act.)F1657.4 GBF16340details
Muse Glimmer 30B30BF1652.9 GBF1637details
granite 4.2 30b29BF1657.1 GBF1635details
Qwen3.8 27B28BF1652.3 GBF1637details
Qwen3.6 27B28BF1651.5 GBF1638details
Qwen3.5 27B28BF1651.5 GBF1638details
gemma 3 27b it27BF1652.2 GBF1638details
Arete 27B27BF1651.5 GBF1638details
MiniMax H3 Image PE27BF1652.3 GBF1637details
Trinity Mini26B (2.0B act.)F1649.7 GBF16519details
GLiDE 27B26BF1650.0 GBF1639details
diffusiongemma 26B A4B it26B (4.0B act.)F1648.5 GBF16260details
gemma 4 26B A4B it26B (4.0B act.)F1648.4 GBF16260details
LFM2.5 24B A2B24B (2.0B act.)Q8_024.5 GBQ8_0959details
ERNIE 4.5 21B A3B PT22BF1641.7 GBF1647details
ERNIE 4.5 21B A3B Thinking22BF1641.7 GBF1647details
gpt oss safeguard 20b22B (4.8B act.)F1613.8 GBF16664details
gpt oss 20b21B (4.2B act.)F1613.8 GBF16737details
Groot 19B A3B 26102319B (3.0B act.)F1636.4 GBF16348details
Kimi VL A3B Thinking 250616B (3.0B act.)F1660.4 GBF16175details
Kimi VL A3B Instruct16B (3.0B act.)F1631.5 GBF16340details
Kimi VL A3B Thinking16B (3.0B act.)F1631.5 GBF16340details
deepseek moe 16b base16B (2.8B act.)F1632.8 GBF16360details
LLaDA2.1 mini16B (1.4B act.)F1631.3 GBF16711details
Moonlight 16B A3B Instruct16B (3.0B act.)F1630.7 GBF16340details
DeepSeek Coder V2 Lite Instruct16B (2.7B act.)F1630.1 GBF16383details
DeepSeek V2 Lite16B (2.7B act.)F1630.1 GBF16383details
DeepSeek V2 Lite Chat16B (2.7B act.)F1630.1 GBF16383details
Qwen2.5 Coder 14B Instruct15BF1629.7 GBF1669details
DeepSeek R1 Distill Qwen 14B15BF1629.7 GBF1669details
Qwen3 14B15BF1629.4 GBF1669details
phi 415BF1629.5 GBF1670details
gemma 3 12b it12BF1623.6 GBF1687details
Llama Guard 4 12B12BF1624.6 GBF1685details
Llama 3.2 11B Vision Instruct11BF1620.1 GBF16104details
Llama 3.2 11B Vision11BF1621.7 GBF1696details
GLM 4.6V Flash10BF1618.5 GBF16108details
GLM 4.1V 9B Thinking10BF1618.5 GBF16108details
Step3 VL 10B10BF1632.3 GBF1662details
SenseNova Loop 9B Preview9.7BF1617.8 GBF16114details
Qwen3.5 9B9.7BF1617.8 GBF16114details
Ornith 1.5 9B9.7BF1617.8 GBF16114details
Qwen3.5 9B Base9.7BF1617.8 GBF16114details
MiMo V2.6 Distill Qwen 9B9.4BF1617.8 GBF16114details
glm 4 9b9.4BF1623.2 GBF16108details
internlm3 8b instruct8.8BF1617.4 GBF16116details
granite 4.2 8b8.8BF1618.2 GBF16116details
granite 4.1 8b8.8BF1618.2 GBF16116details
Qwen3 VL 8B Instruct8.8BF1617.1 GBF16124details
Molmo2 8B8.7BF1617.1 GBF16124details
LFM2.5 8B A1B8.5B (1.0B act.)F1632.3 GBF16509details
Qwen2.5 VL 7B Instruct8.3BF1615.3 GBF16134details
UI TARS 1.5 7B8.3BF1615.3 GBF16134details
olmOCR 2 7B 10258.3BF1615.3 GBF16134details
olmOCR 7B 0225 preview8.3BF1616.6 GBF16123details
UI TARS 7B DPO8.3BF1615.3 GBF16134details
Qwen3 8B8.2BF1617.1 GBF16124details
DeepSeek R1 0528 Qwen3 8B8.2BF1617.1 GBF16124details
granite 3.3 8b instruct8.2BF1617.0 GBF16125details
granite 3.1 8b instruct8.2BF1617.0 GBF16125details
granite 3.2 8b instruct8.2BF1617.0 GBF16125details
Apertus 8B Instruct 25098.1BF1616.7 GBF16127details
Llama 3.1 8B Instruct8.0BF1616.6 GBF16127details
Llama 3.1 8B8.0BF1616.6 GBF16127details
Hermes 3 Llama 3.1 8B8.0BF1616.6 GBF16127details
DeepSeek R1 Distill Llama 8B8.0BF1631.6 GBF1663details
Meta Llama 3 8B Instruct8.0BF1616.6 GBF16127details
Meta Llama Guard 2 8B8.0BF1616.6 GBF16127details
Meta Llama 3.1 8B Instruct8.0BF1616.6 GBF16127details
Ling 3.0 tiny7.9B (1.4B act.)F1615.6 GBF16737details
MiMo 7B RL7.8BF1616.4 GBF16130details
internlm2 5 7b chat7.7BF1616.0 GBF16132details
internlm2 chat 7b7.7BF1616.0 GBF16132details
Qwen2.5 7B Instruct7.6BF1615.3 GBF16134details
Qwen2.5 Coder 7B Instruct7.6BF1615.3 GBF16134details
DeepSeek R1 Distill Qwen 7B7.6BF1615.3 GBF16134details
Falcon H1R 7B7.6BF1615.1 GBF16134details
Olmo 3 7B Instruct7.3BF1616.7 GBF16140details
Olmo 3 1025 7B7.3BF1616.7 GBF16140details
Olmo 3 7B Think7.3BF1616.7 GBF16140details
Mistral 7B v0.17.2BF1614.9 GBF16141details
Mistral 7B Instruct v0.27.2BF1614.9 GBF16141details
Mistral 7B Instruct v0.17.2BF1614.9 GBF16141details
falcon 7b7.2BF1618.4 GBF16141details
falcon 7b instruct7.2BF1618.4 GBF16141details
granite 4.0 h tiny6.9B (1.5B act.)F1613.6 GBF16695details
OLMoE 1B 7B 09246.9B (1.3B act.)F1614.4 GBF16795details
deepseek coder 7b instruct v1.56.9BF1617.2 GBF16148details
deepseek coder 6.7b instruct6.7BF1617.0 GBF16151details
deepseek coder 6.7b base6.7BF1617.0 GBF16151details
Llama 2 7b hf6.7BF1617.0 GBF16151details
Yi 6B 200K6.1BF1612.3 GBF16168details
Yi 6B6.1BF1612.3 GBF16168details
Molmo2 4B4.9BF1610.0 GBF16231details
WeVisOne VL 4B4.8BF1610.0 GBF16231details
Mage VL Cowork4.7BF1610.0 GBF16231details
Qianfan OCR4.7BF1610.1 GBF16230details
Qwen3.5 4B4.7BF169.0 GBF16242details
S5 nano4.7BF169.2 GBF16235details
LightOnOCR 3 4B4.5BF169.0 GBF16242details
SingVL S1 4B4.5BF169.0 GBF16242details
Intern Decision 4B4.5BF169.0 GBF16242details
Qwen3 VL 4B Instruct4.4BF169.3 GBF16253details
gemma 3 4b it4.3BF168.4 GBF16262details
medgemma 1.5 4b it4.3BF168.4 GBF16262details
medgemma 4b it4.3BF168.4 GBF16262details
gemma 3 4b pt4.3BF169.2 GBF16237details
Phi 3.5 vision instruct4.1BF1611.2 GBF16246details
Qwen3 4B4.0BF169.3 GBF16253details
Qwen3 4B Instruct 25074.0BF169.3 GBF16253details
granite vision 4.1 4b4.0BF167.6 GBF16299details
Phi 4 mini instruct3.8BF168.9 GBF16265details
Phi 3 mini 4k instruct3.8BF1610.6 GBF16267details
Phi 3.5 mini instruct3.8BF1610.6 GBF16267details
Phi 3 mini 128k instruct3.8BF1610.6 GBF16267details
Qwen2.5 VL 3B Instruct3.8BF166.7 GBF16330details
granite 4.2 3b3.7BF168.0 GBF16278details
granite 4.1 3b llmon preview3.7BF168.0 GBF16279details
Raon 1 nano3.5BF167.0 GBF16325details
granite 4.1 3b3.4BF167.6 GBF16299details
granite 4.0 micro3.4BF167.6 GBF16299details
DeepSeek OCR 23.4B (1.2B act.)F166.6 GBF16985details
deepseek vl2 tiny3.4B (1.2B act.)F167.4 GBF16867details
Unlimited OCR 23.3B (1.1B act.)F166.6 GBF161009details
Unlimited OCR3.3B (1.1B act.)F166.6 GBF161015details
DeepSeek OCR3.3B (1.1B act.)F166.6 GBF161015details
granite 3.1 3b a800m instruct3.3B (883M act.)F167.1 GBF161154details
Llama 3.2 3B Instruct3.2BF1613.5 GBF16158details
Llama 3.2 3B3.2BF167.5 GBF16317details
Hermes 3 Llama 3.2 3B3.2BF167.5 GBF16317details
d1 3B3.1BF1610.8 GBF16189details
LFM2.5 VL 3B3.1BF1610.8 GBF16189details
Qwen2.5 3B Instruct3.1BF166.7 GBF16330details
SmolLM3 3B3.1BF166.9 GBF16331details
SmolLM3 3B Base3.1BF166.9 GBF16332details
phi 22.8BF168.2 GBF16367details
LFM2.5 2.6B2.7BF1610.8 GBF16189details
granite 3.3 2b instruct2.5BF165.8 GBF16402details
MiniCPM5 2B2.5BF165.7 GBF16405details
WeVisOne VL 2B2.4BF165.4 GBF16500details
Qwen3.5 2B2.3BF164.5 GBF16540details
SmolVLM Instruct2.2BF165.4 GBF16562details
Qwen3 1.7B2.0BF164.8 GBF16591details
Qwen3.8 27B DSpine b162.0BF164.8 GBF16502details
d1 2B1.9BF167.2 GBF16294details
LFM3 VL 2B1.9BF167.2 GBF16294details
DeepSeek R1 Distill Qwen 1.5B1.8BF164.2 GBF16573details
LFM3 1.7B1.7BF167.2 GBF16294details
LFM3 1.7B Base1.7BF164.0 GBF16590details
SmolLM2 1.7B Instruct1.7BF165.2 GBF16595details
SmolLM2 1.7B1.7BF165.2 GBF16596details
SmolLM 1.7B1.7BF165.2 GBF16595details
granite 4.0 1b1.6BF164.3 GBF16624details
Qwen2.5 1.5B Instruct1.5BF163.8 GBF16660details
OLMo 2 0425 1B1.5BF164.4 GBF16686details
deepseek coder 1.3b instruct1.3BF164.5 GBF16757details
Llama 3.2 1B1.2BF163.2 GBF16825details
Llama 3.2 1B Instruct1.2BF165.5 GBF16411details
LFM2.5 1.2B Instruct1.2BF165.0 GBF16435details
HunyuanOCR1.1BF163.5 GBF16910details
PaddleOCR VL 1.6959MF162.5 GBF161064details
Qwen3.5 0.8B873MF162.4 GBF161344details
LightOnOCR 3 0.8B853MF162.4 GBF161344details
FastVLM 0.5B759MF162.2 GBF161344details
Qwen3 0.6B752MF162.7 GBF161702details
GOT OCR2 0716MF162.8 GBF161424details
Nemotron 3.5 Parse631MF161.9 GBF161560details
GOT OCR 2.0 hf561MF162.5 GBF161819details
SmolVLM2 500M Video Instruct507MF161.6 GBF162485details
SmolVLM 500M Instruct507MF161.6 GBF162485details
Qwen2.5 0.5B Instruct494MF161.7 GBF162064details
LFM2.5 VL 450M449MF161.9 GBF161433details
SmolLM2 360M Instruct362MF161.5 GBF162810details
SmolLM2 360M362MF161.5 GBF162810details
LFM2.5 350M354MF161.9 GBF161433details
LFM2 350M354MF161.4 GBF162876details

Questions

What is the biggest model that runs on the A100 80GB?

Of the models we track, the largest that fits at 8K context is Step 3.7 Flash (201B) at Q2_K, needing about 74.5 GB.

How much of the A100 80GB's memory can a model use?

About 76.0 GB: we assume 95% of the 80 GB is usable after driver and display overhead.

Spec source: NVIDIA specifications. Model data: Hugging Face Hub API. Speed is a bandwidth ceiling (bandwidth ÷ bytes read per token), not a benchmark. Methodology · Report an error.