RunsOnMy ModelsGPUs & MacsAPI pricesVRAM calculatorNew

What AI models run on the H200?

The H200 has 141 GB of VRAM (4,800 GB/s). About 134 GB is usable for a model. 243 of 285 open models we track fit at 8K context.

243models fit (8K ctx)
241still fit at 32K ctx
358Blargest that fits
4800 GB/smemory bandwidth

Models that fit, largest first

ModelParamsBest quant (8K)Needs32K contextMax tok/s (est.)
GLM 4.5358B (36B act.)Q2_K126 GBno369details
GLM 4.7358B (36B act.)Q2_K126 GBno369details
GLM 4.6357B (34B act.)Q2_K125 GBQ2_K385details
GLM 5.3 Flash321B (18B act.)Q2_K113 GBQ2_K698details
Nex N2.6321B (18B act.)Q2_K118 GBQ2_K670details
MiMo V2.6 Flash RL311B (6.2B act.)Q2_K110 GBQ2_K—details
MiMo V2.5311B (17B act.)Q2_K110 GBQ2_K735details
MiMo V2.6 Flash MOPD311B (6.2B act.)Q2_K110 GBQ2_K—details
MiMo V2 Flash310B (16B act.)Q2_K106 GBQ2_K806details
DeepSeek V4 Flash Vision Exp305B (15B act.)Q2_K108 GBQ2_K840details
DeepSeek V4 Flash 0731304B (27B act.)Q2_K108 GBQ2_K468details
DeepSeek V4 Flash DSpark304B (27B act.)Q2_K108 GBQ2_K468details
Hy3299B (21B act.)Q2_K102 GBQ2_K641details
Hy3 preview299B (21B act.)Q2_K107 GBQ2_K611details
DeepSeek V4 Flash291B (14B act.)Q2_K103 GBQ2_K906details
ERNIE 5.5 Flash Base286B (5.7B act.)Q2_K101 GBQ2_K—details
Intern S1241B (28B act.)Q3_K_M112 GBQ3_K_M353details
DeepSeek V2 Chat236B (21B act.)Q3_K_M108 GBQ3_K_M459details
MiniMax M2.5229B (11B act.)Q3_K_M104 GBQ3_K_M911details
MiniMax M2.7229B (11B act.)Q3_K_M97.0 GBQ3_K_M983details
MiniMax M2229B (11B act.)Q3_K_M104 GBQ3_K_M911details
MiniMax M2.1229B (11B act.)Q3_K_M104 GBQ3_K_M911details
MiniMax M3.1 Flash220B (4.4B act.)Q3_K_M102 GBQ3_K_M—details
Step 3.7 Flash201BQ4_K_M119 GBQ3_K_M39details
Step 3.5 Flash199BQ4_K_M116 GBQ3_K_M40details
Qwen3.8 Flash Next180B (62B act.)Q5_K_M121 GBQ4_K_M109details
Ling 3.0 flash Sante127B (5.6B act.)Q6_K105 GBQ6_K980details
Ling 3.0 flash127B (5.6B act.)Q6_K103 GBQ6_K999details
Ornith 1.5 122B125B (13B act.)Q6_K96.6 GBQ6_K459details
gpt oss 120b117B (5.8B act.)F1661.9 GBF161492details
gpt oss safeguard 120b117B (5.8B act.)Q8_060.1 GBQ8_01538details
GLM 4.5 Air110B (15B act.)Q8_0112 GBQ8_0302details
Llama 4 Scout 17B 16E Instruct109B (18B act.)Q8_0109 GBQ8_0252details
Llama 4 Scout 17B 16E109B (18B act.)Q8_0110 GBQ8_0250details
GLM 4.5V108B (14B act.)Q8_0108 GBQ8_0319details
GLM 4.6V108B (14B act.)Q8_0108 GBQ8_0319details
Hunyuan A13B Instruct80B (13B act.)Q8_081.2 GBQ8_0347details
Kimi Dev 72B73BQ8_075.1 GBQ8_062details
Llama 3.3 70B Instruct71BQ8_073.0 GBQ8_064details
DeepSeek R1 Distill Llama 70B71BQ8_073.0 GBQ8_064details
Llama 3.1 70B71BQ8_073.0 GBQ8_064details
Kimi Linear 48B A3B Instruct49B (3.0B act.)F1692.5 GBF16800details
Nous Hermes 2 Mixtral 8x7B DPO47B (13B act.)F1688.5 GBF16186details
falcon 40b42BF1679.4 GBF1657details
flint40B (1.3B act.)F1675.1 GBF161920details
Maxwell V1 Flash37B (2.3B act.)F1669.7 GBF161057details
Hermes 4.3 36B36BF1670.0 GBF1666details
Seed OSS 36B Instruct36BF1670.0 GBF1666details
Qwen3.6 35B A3B36B (3.0B act.)F1667.2 GBF16809details
Ornith 1.5 35B A3B36B (3.0B act.)F1667.2 GBF16809details
Qwen3.5 35B A3B36B (3.0B act.)F1665.6 GBF16829details
Axion 1.0 35B35B (3.9B act.)F1665.6 GBF16623details
Qwen AgentWorld 35B A3B35B (3.0B act.)F1665.6 GBF16799details
Yi 34B 200K34BF1666.5 GBF1670details
Yi 34B34BF1666.5 GBF1670details
Yi 34B Chat34BF1666.5 GBF1670details
Yi 1.5 34B Chat34BF1666.5 GBF1670details
Nous Hermes 2 Yi 34B34BF1666.5 GBF1670details
DeepSeek R1 Distill Qwen 32B33BF1663.7 GBF1673details
Qwen3 32B33BF1663.7 GBF1673details
GLM Z1 32B 041433BF1661.8 GBF1674details
GLM 4 32B 041433BF1661.8 GBF1674details
granite 4.0 h small32B (8.8B act.)F1660.7 GBF16272details
gemma 4 31B it31BF1660.1 GBF1678details
GLM 4.7 Flash31B (4.6B act.)F1656.9 GBF16543details
Qwen3 Coder 30B A3B Instruct31B (3.0B act.)F1658.3 GBF16800details
North Mini Code 1.030B (3.3B act.)F1658.2 GBF16726details
Hy Memory 30B A3B30B (3.0B act.)F1657.4 GBF16800details
Muse Glimmer 30B30BF1652.9 GBF1686details
granite 4.2 30b29BF1657.1 GBF1682details
Qwen3.8 27B28BF1652.3 GBF1688details
Qwen3.6 27B28BF1651.5 GBF1689details
Qwen3.5 27B28BF1651.5 GBF1689details
gemma 3 27b it27BF1652.2 GBF1689details
Arete 27B27BF1651.5 GBF1689details
MiniMax H3 Image PE27BF1652.3 GBF1688details
Trinity Mini26B (2.0B act.)F1649.7 GBF161222details
GLiDE 27B26BF1650.0 GBF1692details
diffusiongemma 26B A4B it26B (4.0B act.)F1648.5 GBF16613details
gemma 4 26B A4B it26B (4.0B act.)F1648.4 GBF16613details
LFM2.5 24B A2B24B (2.0B act.)F1690.1 GBF16600details
ERNIE 4.5 21B A3B PT22BF1641.7 GBF16110details
ERNIE 4.5 21B A3B Thinking22BF1641.7 GBF16110details
gpt oss safeguard 20b22B (4.8B act.)F1613.8 GBF161563details
gpt oss 20b21B (4.2B act.)F1613.8 GBF161736details
Groot 19B A3B 26102319B (3.0B act.)F1636.4 GBF16818details
Kimi VL A3B Thinking 250616B (3.0B act.)F1660.4 GBF16411details
Kimi VL A3B Instruct16B (3.0B act.)F1631.5 GBF16800details
Kimi VL A3B Thinking16B (3.0B act.)F1631.5 GBF16800details
deepseek moe 16b base16B (2.8B act.)F1632.8 GBF16848details
LLaDA2.1 mini16B (1.4B act.)F1631.3 GBF161675details
Moonlight 16B A3B Instruct16B (3.0B act.)F1630.7 GBF16800details
DeepSeek Coder V2 Lite Instruct16B (2.7B act.)F1630.1 GBF16902details
DeepSeek V2 Lite16B (2.7B act.)F1630.1 GBF16902details
DeepSeek V2 Lite Chat16B (2.7B act.)F1630.1 GBF16902details
Qwen2.5 Coder 14B Instruct15BF1629.7 GBF16162details
DeepSeek R1 Distill Qwen 14B15BF1629.7 GBF16162details
Qwen3 14B15BF1629.4 GBF16163details
phi 415BF1629.5 GBF16164details
gemma 3 12b it12BF1623.6 GBF16204details
Llama Guard 4 12B12BF1624.6 GBF16200details
Llama 3.2 11B Vision Instruct11BF1620.1 GBF16245details
Llama 3.2 11B Vision11BF1621.7 GBF16226details
GLM 4.6V Flash10BF1618.5 GBF16255details
GLM 4.1V 9B Thinking10BF1618.5 GBF16255details
Step3 VL 10B10BF1632.3 GBF16146details
SenseNova Loop 9B Preview9.7BF1617.8 GBF16268details
Qwen3.5 9B9.7BF1617.8 GBF16268details
Ornith 1.5 9B9.7BF1617.8 GBF16268details
Qwen3.5 9B Base9.7BF1617.8 GBF16268details
MiMo V2.6 Distill Qwen 9B9.4BF1617.8 GBF16268details
glm 4 9b9.4BF1623.2 GBF16255details
internlm3 8b instruct8.8BF1617.4 GBF16273details
granite 4.2 8b8.8BF1618.2 GBF16273details
granite 4.1 8b8.8BF1618.2 GBF16273details
Qwen3 VL 8B Instruct8.8BF1617.1 GBF16293details
Molmo2 8B8.7BF1617.1 GBF16293details
LFM2.5 8B A1B8.5B (1.0B act.)F1632.3 GBF161199details
Qwen2.5 VL 7B Instruct8.3BF1615.3 GBF16315details
UI TARS 1.5 7B8.3BF1615.3 GBF16315details
olmOCR 2 7B 10258.3BF1615.3 GBF16315details
olmOCR 7B 0225 preview8.3BF1616.6 GBF16289details
UI TARS 7B DPO8.3BF1615.3 GBF16315details
Qwen3 8B8.2BF1617.1 GBF16293details
DeepSeek R1 0528 Qwen3 8B8.2BF1617.1 GBF16293details
granite 3.3 8b instruct8.2BF1617.0 GBF16294details
granite 3.1 8b instruct8.2BF1617.0 GBF16294details
granite 3.2 8b instruct8.2BF1617.0 GBF16294details
Apertus 8B Instruct 25098.1BF1616.7 GBF16298details
Llama 3.1 8B Instruct8.0BF1616.6 GBF16299details
Llama 3.1 8B8.0BF1616.6 GBF16299details
Hermes 3 Llama 3.1 8B8.0BF1616.6 GBF16299details
DeepSeek R1 Distill Llama 8B8.0BF1631.6 GBF16149details
Meta Llama 3 8B Instruct8.0BF1616.6 GBF16299details
Meta Llama Guard 2 8B8.0BF1616.6 GBF16299details
Meta Llama 3.1 8B Instruct8.0BF1616.6 GBF16299details
Ling 3.0 tiny7.9B (1.4B act.)F1615.6 GBF161735details
MiMo 7B RL7.8BF1616.4 GBF16306details
internlm2 5 7b chat7.7BF1616.0 GBF16310details
internlm2 chat 7b7.7BF1616.0 GBF16310details
Qwen2.5 7B Instruct7.6BF1615.3 GBF16315details
Qwen2.5 Coder 7B Instruct7.6BF1615.3 GBF16315details
DeepSeek R1 Distill Qwen 7B7.6BF1615.3 GBF16315details
Falcon H1R 7B7.6BF1615.1 GBF16316details
Olmo 3 7B Instruct7.3BF1616.7 GBF16329details
Olmo 3 1025 7B7.3BF1616.7 GBF16329details
Olmo 3 7B Think7.3BF1616.7 GBF16329details
Mistral 7B v0.17.2BF1614.9 GBF16331details
Mistral 7B Instruct v0.27.2BF1614.9 GBF16331details
Mistral 7B Instruct v0.17.2BF1614.9 GBF16331details
falcon 7b7.2BF1618.4 GBF16333details
falcon 7b instruct7.2BF1618.4 GBF16333details
granite 4.0 h tiny6.9B (1.5B act.)F1613.6 GBF161636details
OLMoE 1B 7B 09246.9B (1.3B act.)F1614.4 GBF161871details
deepseek coder 7b instruct v1.56.9BF1617.2 GBF16347details
deepseek coder 6.7b instruct6.7BF1617.0 GBF16356details
deepseek coder 6.7b base6.7BF1617.0 GBF16356details
Llama 2 7b hf6.7BF1617.0 GBF16356details
Yi 6B 200K6.1BF1612.3 GBF16396details
Yi 6B6.1BF1612.3 GBF16396details
Molmo2 4B4.9BF1610.0 GBF16544details
WeVisOne VL 4B4.8BF1610.0 GBF16543details
Mage VL Cowork4.7BF1610.0 GBF16544details
Qianfan OCR4.7BF1610.1 GBF16543details
Qwen3.5 4B4.7BF169.0 GBF16570details
S5 nano4.7BF169.2 GBF16554details
LightOnOCR 3 4B4.5BF169.0 GBF16570details
SingVL S1 4B4.5BF169.0 GBF16570details
Intern Decision 4B4.5BF169.0 GBF16570details
Qwen3 VL 4B Instruct4.4BF169.3 GBF16596details
gemma 3 4b it4.3BF168.4 GBF16618details
medgemma 1.5 4b it4.3BF168.4 GBF16618details
medgemma 4b it4.3BF168.4 GBF16618details
gemma 3 4b pt4.3BF169.2 GBF16558details
Phi 3.5 vision instruct4.1BF1611.2 GBF16579details
Qwen3 4B4.0BF169.3 GBF16597details
Qwen3 4B Instruct 25074.0BF169.3 GBF16596details
granite vision 4.1 4b4.0BF167.6 GBF16705details
Phi 4 mini instruct3.8BF168.9 GBF16625details
Phi 3 mini 4k instruct3.8BF1610.6 GBF16628details
Phi 3.5 mini instruct3.8BF1610.6 GBF16628details
Phi 3 mini 128k instruct3.8BF1610.6 GBF16628details
Qwen2.5 VL 3B Instruct3.8BF166.7 GBF16777details
granite 4.2 3b3.7BF168.0 GBF16655details
granite 4.1 3b llmon preview3.7BF168.0 GBF16656details
Raon 1 nano3.5BF167.0 GBF16765details
granite 4.1 3b3.4BF167.6 GBF16705details
granite 4.0 micro3.4BF167.6 GBF16705details
DeepSeek OCR 23.4B (1.2B act.)F166.6 GBF162318details
deepseek vl2 tiny3.4B (1.2B act.)F167.4 GBF162042details
Unlimited OCR 23.3B (1.1B act.)F166.6 GBF162374details
Unlimited OCR3.3B (1.1B act.)F166.6 GBF162388details
DeepSeek OCR3.3B (1.1B act.)F166.6 GBF162388details
granite 3.1 3b a800m instruct3.3B (883M act.)F167.1 GBF162716details
Llama 3.2 3B Instruct3.2BF1613.5 GBF16373details
Llama 3.2 3B3.2BF167.5 GBF16747details
Hermes 3 Llama 3.2 3B3.2BF167.5 GBF16746details
d1 3B3.1BF1610.8 GBF16444details
LFM2.5 VL 3B3.1BF1610.8 GBF16444details
Qwen2.5 3B Instruct3.1BF166.7 GBF16778details
SmolLM3 3B3.1BF166.9 GBF16779details
SmolLM3 3B Base3.1BF166.9 GBF16780details
phi 22.8BF168.2 GBF16863details
LFM2.5 2.6B2.7BF1610.8 GBF16444details
granite 3.3 2b instruct2.5BF165.8 GBF16947details
MiniCPM5 2B2.5BF165.7 GBF16953details
WeVisOne VL 2B2.4BF165.4 GBF161178details
Qwen3.5 2B2.3BF164.5 GBF161271details
SmolVLM Instruct2.2BF165.4 GBF161324details
Qwen3 1.7B2.0BF164.8 GBF161392details
Qwen3.8 27B DSpine b162.0BF164.8 GBF161182details
d1 2B1.9BF167.2 GBF16692details
LFM3 VL 2B1.9BF167.2 GBF16692details
DeepSeek R1 Distill Qwen 1.5B1.8BF164.2 GBF161348details
LFM3 1.7B1.7BF167.2 GBF16692details
LFM3 1.7B Base1.7BF164.0 GBF161388details
SmolLM2 1.7B Instruct1.7BF165.2 GBF161402details
SmolLM2 1.7B1.7BF165.2 GBF161402details
SmolLM 1.7B1.7BF165.2 GBF161402details
granite 4.0 1b1.6BF164.3 GBF161469details
Qwen2.5 1.5B Instruct1.5BF163.8 GBF161555details
OLMo 2 0425 1B1.5BF164.4 GBF161614details
deepseek coder 1.3b instruct1.3BF164.5 GBF161782details
Llama 3.2 1B1.2BF163.2 GBF161942details
Llama 3.2 1B Instruct1.2BF165.5 GBF16968details
LFM2.5 1.2B Instruct1.2BF165.0 GBF161024details
HunyuanOCR1.1BF163.5 GBF162143details
PaddleOCR VL 1.6959MF162.5 GBF162504details
Qwen3.5 0.8B873MF162.4 GBF163165details
LightOnOCR 3 0.8B853MF162.4 GBF163165details
FastVLM 0.5B759MF162.2 GBF163163details
Qwen3 0.6B752MF162.7 GBF164006details
GOT OCR2 0716MF162.8 GBF163352details
Nemotron 3.5 Parse631MF161.9 GBF163673details
GOT OCR 2.0 hf561MF162.5 GBF164282details
SmolVLM2 500M Video Instruct507MF161.6 GBF165851details
SmolVLM 500M Instruct507MF161.6 GBF165851details
Qwen2.5 0.5B Instruct494MF161.7 GBF164858details
LFM2.5 VL 450M449MF161.9 GBF163373details
SmolLM2 360M Instruct362MF161.5 GBF166616details
SmolLM2 360M362MF161.5 GBF166616details
LFM2.5 350M354MF161.9 GBF163373details
LFM2 350M354MF161.4 GBF166770details

Just too big at 4-bit

These need up to 35% more memory than the H200 has at Q4_K_M and 8K context. A smaller quant or shorter context may work.

ModelQ4_K_M needsShort by
DeepSeek V2 Chat135.3 GB1.4 GB
Intern S1139.2 GB5.2 GB
ERNIE 5.5 Flash Base164.2 GB30.2 GB
DeepSeek V4 Flash167.0 GB33.1 GB
Hy3172.8 GB38.8 GB
Hy3 preview173.2 GB39.3 GB
DeepSeek V4 Flash 0731174.6 GB40.6 GB
DeepSeek V4 Flash DSpark174.6 GB40.6 GB
DeepSeek V4 Flash Vision Exp174.8 GB40.9 GB
MiMo V2 Flash175.8 GB41.8 GB
MiMo V2.6 Flash RL178.7 GB44.8 GB
MiMo V2.6 Flash MOPD178.7 GB44.8 GB
MiMo V2.5179.6 GB45.7 GB

Questions

What is the biggest model that runs on the H200?

Of the models we track, the largest that fits at 8K context is GLM 4.5 (358B) at Q2_K, needing about 126 GB.

How much of the H200's memory can a model use?

About 133.9 GB: we assume 95% of the 141 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.