Skip to content

What can the 8× H200 NVL run?

What fits on the 8× H200 NVL
Text models: 38Visual models: 0Strongest first
Memory
1128 GB
Bandwidth
4,800 GB/s each
Maker
NVIDIA

Open the calculator on this machineH200 NVL in our inventory

Text models that run on it

Text models that run on the 8× H200 NVL, best first
ModelBest fitContextSpeedStrongest at
Kimi K32-bit1M58 tok/sChat, Coding, Agents, Docs
MiMo V2.6 ProFP4 + FP81M71 tok/sChat, Coding, Agents, Docs
DeepSeek V4.1 FlashFP4 + FP81M135 tok/sChat, Coding, Agents, Docs
GLM-5.3FP81M60 tok/sChat, Coding, Agents, Docs
GLM-5.3 FlashFP81M122 tok/sChat, Coding, Agents, Docs
Qwen3.8 2.4T A95B2-bit256K61 tok/sChat, Coding, Agents
Muse Spark Open 816B A42B8-bit1M59 tok/sChat, Docs
MiniMax M38-bit1M92 tok/sDocs
DeepSeek V4 ProFP4 + FP81M64 tok/sChat, Coding, Agents, Docs
Kimi K2.6INT4256K58 tok/sChat, Coding, Agents, Docs
Hy38-bit256K87 tok/sChat, Coding, Agents, Docs
Nemotron 3 Ultra 550B A55B8-bit256K42 tok/sChat, Coding, Docs
Qwen3.5 397B A17B8-bit256K111 tok/sAgents, Docs
DeepSeek V3.2FP8160K67 tok/sChat, Coding, Agents, Docs
Inkling8-bit4K59 tok/sText and images
DeepSeek R1FP8160K64 tok/sNot rated
DeepSeek V3FP8160K64 tok/sNot rated
GLM 58-bit198K55 tok/sNot rated
GLM 5.18-bit198K55 tok/sNot rated
GLM 5.28-bit1M56 tok/sNot rated
Kimi K2 InstructFP8128K62 tok/sNot rated
Ornith 1.5 397B8-bit256K88 tok/sNot rated
Ornith 1.0 397B8-bit256K108 tok/sNot rated
DeepSeek V3 0324FP8160K64 tok/sNot rated
DeepSeek R1 0528FP8160K64 tok/sNot rated
GLM 4.58-bit128K60 tok/sNot rated
MiMo V2.5FP81M123 tok/sNot rated
GLM 5.3 Flash BF168-bit1M117 tok/sText and images
Hy3 Preview8-bit256K88 tok/sNot rated
Qwen3 Coder 480B A35B Instruct8-bit256K68 tok/sCoding
GLM 4.78-bit198K60 tok/sNot rated
GLM 5.3 BF168-bit1M56 tok/sNot rated
GLM 4.68-bit198K61 tok/sNot rated
MiMo V2 FlashFP8256K128 tok/sNot rated
K EXAONE 2.0 750B A37B8-bit256K61 tok/sNot rated
Beam 1.0 501B A23B8-bit1M96 tok/sNot rated
Hy4 Preview8-bit1M49 tok/sNot rated
K2 Horizon 375B A23B8-bit512K91 tok/sNot rated

Best fit is the highest quality, up to 8-bit, that leaves the use case’s context. Ranked by our use-case scores, then by Hugging Face downloads. Speeds are rough, plus or minus 25%.

Questions

What is the best LLM for the 8× H200 NVL?

Of the open models we track, the strongest that run well on the 8× H200 NVL are Kimi K3, MiMo V2.6 Pro, DeepSeek V4.1 Flash. 38 text models fit it in all.

Calculated from published specs, not benchmarked. All machines