What can the 4× RTX PRO 6000 Blackwell run?
| Text models: 69 | Visual models: 0 | Strongest 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
| Model | Best fit | Context | Speed | Strongest at |
|---|---|---|---|---|
| MiMo V2.6 Pro | 2-bit | 80K | 60 tok/s | Chat, Coding, Agents, Docs |
| DeepSeek V4.1 Flash | 3-bit | 1M | 111 tok/s | Chat, Coding, Agents, Docs |
| GLM-5.3 | 3-bit | 192K | 49 tok/s | Chat, Coding, Agents, Docs |
| GLM-5.3 Flash | FP8 | 1M | 57 tok/s | Chat, Coding, Agents, Docs |
| Qwen3.8 Flash-Next | 8-bit | 256K | 124 tok/s | Chat, Coding, Agents |
| Muse Spark Open 816B A42B | 2-bit | 1M | 61 tok/s | Chat, Docs |
| MiMo V2.6 Flash | FP4 + FP8 | 1M | 87 tok/s | Coding, Agents |
| MiniMax M3 | 6-bit | 288K | 54 tok/s | Docs |
| DeepSeek V4 Flash | FP4 + FP8 | 1M | 57 tok/s | Docs |
| DeepSeek V4 Pro | 1.6-bit | 1M | 85 tok/s | Chat, Coding, Agents, Docs |
| Kimi K2.6 | 2-bit | 44K | 76 tok/s | Chat, Coding, Agents, Docs |
| Hy3 | 8-bit | 224K | 44 tok/s | Chat, Coding, Agents, Docs |
| Nemotron 3 Ultra 550B A55B | 4-bit | 256K | 30 tok/s | Chat, Coding, Docs |
| Qwen3.5 397B A17B | 6-bit | 256K | 67 tok/s | Agents, Docs |
| MiniMax M2.7 | FP8 | 200K | 85 tok/s | Coding, Agents |
| DeepSeek V3.2 | 3-bit | 160K | 55 tok/s | Chat, Coding, Agents, Docs |
| Mistral Medium 3.5 128B | FP8 | 256K | 9.3 tok/s | Chat, Coding, Agents, Docs |
| Qwen3.5 122B A10B | 8-bit | 256K | 87 tok/s | Chat, Coding, Agents, Docs |
| Nemotron 3 Super 120B A12B | 8-bit | 256K | 64 tok/s | Chat, Coding, Agents, Docs |
| Mistral Small 4 119B | FP8 | 256K | 135 tok/s | Chat, Docs |
| Inkling | 2-bit | 4K | 62 tok/s | Text and images |
| DeepSeek R1 | 3-bit | 160K | 52 tok/s | Not rated |
| DeepSeek V3 | 3-bit | 160K | 52 tok/s | Not rated |
| GLM 5 | 3-bit | 192K | 47 tok/s | Not rated |
| GLM 4.5 Air | 8-bit | 128K | 65 tok/s | Not rated |
| Inkling Small | 8-bit | 4K | 82 tok/s | Text and images |
| MiniMax M2.5 | FP8 | 192K | 77 tok/s | Not rated |
| GLM 5.1 | 3-bit | 192K | 47 tok/s | Not rated |
| Laguna M.1 | 8-bit | 256K | 42 tok/s | Not rated |
| Laguna S 2.1 | 8-bit | 1M | 100 tok/s | Not rated |
| Qwen3 VL 235B A22B Instruct | 8-bit | 256K | 41 tok/s | Text and images |
| GLM 5.2 | 3-bit | 192K | 48 tok/s | Not rated |
| Qwen3 235B A22B | 8-bit | 40K | 41 tok/s | Not rated |
| Ornith 1.5 122B | 8-bit | 256K | 72 tok/s | Not rated |
| Kimi K2 Instruct | 2-bit | 44K | 62 tok/s | Not rated |
| Ornith 1.5 397B | 6-bit | 256K | 51 tok/s | Not rated |
| Ornith 1.0 397B | 6-bit | 256K | 65 tok/s | Not rated |
| Step 3.5 Flash | 8-bit | 256K | 94 tok/s | Not rated |
| DeepSeek V3 0324 | 3-bit | 160K | 52 tok/s | Not rated |
| DeepSeek R1 0528 | 3-bit | 160K | 52 tok/s | Not rated |
| GLM 4.5 | 8-bit | 20K | 28 tok/s | Not rated |
| Llama 4 Scout 17B 16E Instruct | 8-bit | 1.3M | 55 tok/s | Text and images |
| MiniMax M2 | FP8 | 192K | 77 tok/s | Not rated |
| MiMo V2.5 | FP8 | 1M | 59 tok/s | Not rated |
| Solar Open 100B | 8-bit | 128K | 76 tok/s | Not rated |
| GLM 5.3 Flash BF16 | 8-bit | 1M | 55 tok/s | Text and images |
| Qwen3 VL 235B A22B Thinking | 8-bit | 256K | 41 tok/s | Text and images |
| Qwen3 235B A22B Instruct 2507 | 8-bit | 256K | 41 tok/s | Not rated |
| NVIDIA Nemotron 3.5 Super VL 120B A12B BF16 | 8-bit | 1M | 64 tok/s | Text and images |
| Hy3 Preview | 8-bit | 208K | 44 tok/s | Not rated |
| Qwen3 Coder 480B A35B Instruct | 5-bit | 176K | 42 tok/s | Coding |
| GLM 4.7 | 8-bit | 20K | 28 tok/s | Not rated |
| GLM 5.3 BF16 | 3-bit | 192K | 48 tok/s | Not rated |
| GLM 4.6 | 8-bit | 24K | 29 tok/s | Not rated |
| DeepSeek V2 Chat | 8-bit | 160K | 46 tok/s | Not rated |
| MiMo V2 Flash | FP8 | 256K | 62 tok/s | Not rated |
| Le Triomphant ECE TW3 | 8-bit | 32K | 16 tok/s | Not rated |
| TW3 JRGL V2 | 8-bit | 32K | 16 tok/s | Not rated |
| Command A Plus 05 2026 Bf16 | 8-bit | 195K | 43 tok/s | Text and images |
| GLM 4.5V | 8-bit | 64K | 67 tok/s | Text 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