Can I run NVIDIA-Nemotron-3.5-Bolt-74B-A9B-GGUF on an RTX 5090?
Yes, just
Its UD-IQ2_M file (24.6 GB) fits in the 28.8 GB an RTX 5090 can use, with room for ≈ 256K tokens of context. That file is under 3 bits per weight, so expect noticeably lower quality.
Full analysis of NVIDIA-Nemotron-3.5-Bolt-74B-A9B-GGUF
Everything that runs on RTX 5090
Get the UD-IQ2_M file on Hugging Face
ollama run hf.co/unsloth/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-GGUF:UD-IQ2_MFiles and context on RTX 5090
Each file of NVIDIA-Nemotron-3.5-Bolt-74B-A9B-GGUF (and of GGUF versions the site has analyzed), and how much context fits next to it, with the context cache in 16-bit or 8-bit.
| File | Size | On RTX 5090 | Context (16-bit) | Context (8-bit) |
|---|---|---|---|---|
| BF16 | 148 GB | Doesn't fit | — | — |
| Q8_0 | 78.9 GB | Doesn't fit | — | — |
| UD-Q6_K_XL | 72.4 GB | Doesn't fit | — | — |
| UD-Q5_K_XL | 58.9 GB | Doesn't fit | — | — |
| UD-Q4_K_XL | 50.2 GB | Doesn't fit | — | — |
| UD-Q3_K_XL | 39.0 GB | Doesn't fit | — | — |
| UD-IQ3_XXS | 31.6 GB | Doesn't fit | — | — |
| UD-Q2_K_XL | 28.8 GB | Doesn't fit | — | — |
| UD-IQ2_M | 24.6 GB | Tight | 256K | 256K |
| IQ2_XXS | 22.0 GB | Fits | 256K | 256K |
| UD-IQ2_XXS | 22.0 GB | Fits | 256K | 256K |
| UD-IQ1_M | 20.3 GB | Fits | 256K | 256K |
| UD-IQ1_S | 18.5 GB | Fits | 256K | 256K |