Skip to content

What can the 4× RTX PRO 6000 Blackwell run?

What fits on the 4× RTX PRO 6000 Blackwell
Text models: 69Visual models: 0Strongest first
Memory
384 GB
Bandwidth
1,792 GB/s each
Maker
NVIDIA

Open the calculator on this machineRTX PRO 6000 Blackwell in our inventory

Text models that run on it

Text models that run on the 4× RTX PRO 6000 Blackwell, best first
ModelBest fitContextSpeedStrongest at
MiMo V2.6 Pro2-bit80K60 tok/sChat, Coding, Agents, Docs
DeepSeek V4.1 Flash3-bit1M111 tok/sChat, Coding, Agents, Docs
GLM-5.33-bit192K49 tok/sChat, Coding, Agents, Docs
GLM-5.3 FlashFP81M57 tok/sChat, Coding, Agents, Docs
Qwen3.8 Flash-Next8-bit256K124 tok/sChat, Coding, Agents
Muse Spark Open 816B A42B2-bit1M61 tok/sChat, Docs
MiMo V2.6 FlashFP4 + FP81M87 tok/sCoding, Agents
MiniMax M36-bit288K54 tok/sDocs
DeepSeek V4 FlashFP4 + FP81M57 tok/sDocs
DeepSeek V4 Pro1.6-bit1M85 tok/sChat, Coding, Agents, Docs
Kimi K2.62-bit44K76 tok/sChat, Coding, Agents, Docs
Hy38-bit224K44 tok/sChat, Coding, Agents, Docs
Nemotron 3 Ultra 550B A55B4-bit256K30 tok/sChat, Coding, Docs
Qwen3.5 397B A17B6-bit256K67 tok/sAgents, Docs
MiniMax M2.7FP8200K85 tok/sCoding, Agents
DeepSeek V3.23-bit160K55 tok/sChat, Coding, Agents, Docs
Mistral Medium 3.5 128BFP8256K9.3 tok/sChat, Coding, Agents, Docs
Qwen3.5 122B A10B8-bit256K87 tok/sChat, Coding, Agents, Docs
Nemotron 3 Super 120B A12B8-bit256K64 tok/sChat, Coding, Agents, Docs
Mistral Small 4 119BFP8256K135 tok/sChat, Docs
Inkling2-bit4K62 tok/sText and images
DeepSeek R13-bit160K52 tok/sNot rated
DeepSeek V33-bit160K52 tok/sNot rated
GLM 53-bit192K47 tok/sNot rated
GLM 4.5 Air8-bit128K65 tok/sNot rated
Inkling Small8-bit4K82 tok/sText and images
MiniMax M2.5FP8192K77 tok/sNot rated
GLM 5.13-bit192K47 tok/sNot rated
Laguna M.18-bit256K42 tok/sNot rated
Laguna S 2.18-bit1M100 tok/sNot rated
Qwen3 VL 235B A22B Instruct8-bit256K41 tok/sText and images
GLM 5.23-bit192K48 tok/sNot rated
Qwen3 235B A22B8-bit40K41 tok/sNot rated
Ornith 1.5 122B8-bit256K72 tok/sNot rated
Kimi K2 Instruct2-bit44K62 tok/sNot rated
Ornith 1.5 397B6-bit256K51 tok/sNot rated
Ornith 1.0 397B6-bit256K65 tok/sNot rated
Step 3.5 Flash8-bit256K94 tok/sNot rated
DeepSeek V3 03243-bit160K52 tok/sNot rated
DeepSeek R1 05283-bit160K52 tok/sNot rated
GLM 4.58-bit20K28 tok/sNot rated
Llama 4 Scout 17B 16E Instruct8-bit1.3M55 tok/sText and images
MiniMax M2FP8192K77 tok/sNot rated
MiMo V2.5FP81M59 tok/sNot rated
Solar Open 100B8-bit128K76 tok/sNot rated
GLM 5.3 Flash BF168-bit1M55 tok/sText and images
Qwen3 VL 235B A22B Thinking8-bit256K41 tok/sText and images
Qwen3 235B A22B Instruct 25078-bit256K41 tok/sNot rated
NVIDIA Nemotron 3.5 Super VL 120B A12B BF168-bit1M64 tok/sText and images
Hy3 Preview8-bit208K44 tok/sNot rated
Qwen3 Coder 480B A35B Instruct5-bit176K42 tok/sCoding
GLM 4.78-bit20K28 tok/sNot rated
GLM 5.3 BF163-bit192K48 tok/sNot rated
GLM 4.68-bit24K29 tok/sNot rated
DeepSeek V2 Chat8-bit160K46 tok/sNot rated
MiMo V2 FlashFP8256K62 tok/sNot rated
Le Triomphant ECE TW38-bit32K16 tok/sNot rated
TW3 JRGL V28-bit32K16 tok/sNot rated
Command A Plus 05 2026 Bf168-bit195K43 tok/sText and images
GLM 4.5V8-bit64K67 tok/sText and images

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%. The top 60 of 69 are shown.

Only just too big

These need up to a quarter more memory than the 4× RTX PRO 6000 Blackwell has. Spilling into system RAM or a lower precision can get them running, slowly.

Questions

What is the best LLM for the 4× RTX PRO 6000 Blackwell?

Of the open models we track, the strongest that run well on the 4× RTX PRO 6000 Blackwell are MiMo V2.6 Pro, DeepSeek V4.1 Flash, GLM-5.3. 69 text models fit it in all.

Calculated from published specs, not benchmarked. All machines