Can I run Mellum2.1-12B-A2.5B-Thinking on an RTX 5060?
Yes, with a GGUF version
The published weights are too big, but the MXFP4_MOE file from JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF (7.03 GB) fits, with room for ≈ 9K tokens of context.
Full analysis of Mellum2.1-12B-A2.5B-Thinking
Everything that runs on RTX 5060
Get the MXFP4_MOE file on Hugging Face
ollama run hf.co/JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF:MXFP4_MOEFiles and context on RTX 5060
Each file of Mellum2.1-12B-A2.5B-Thinking (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 5060 | Context (16-bit) | Context (8-bit) |
|---|---|---|---|---|
| As published | 24.3 GB | Doesn't fit | — | — |
| Mellum2.1-12B-A2.5B-Thinking-MTP-BF16 | 25.2 GB | Doesn't fit | — | — |
| BF16 | 24.3 GB | Doesn't fit | — | — |
| Mellum2.1-12B-A2.5B-Thinking-MTP-Q8_0 | 13.4 GB | Doesn't fit | — | — |
| Q8_0 | 12.9 GB | Doesn't fit | — | — |
| Mellum2.1-12B-A2.5B-Thinking-MTP-Q6_K | 11.3 GB | Doesn't fit | — | — |
| Q6_K | 10.9 GB | Doesn't fit | — | — |
| Mellum2.1-12B-A2.5B-Thinking-MTP-Q4_K_M | 8.33 GB | Doesn't fit | — | — |
| Q4_K_M | 8.07 GB | Doesn't fit | — | — |
| Mellum2.1-12B-A2.5B-Thinking-MTP-MXFP4_MOE | 7.27 GB | Doesn't fit | — | — |
| MXFP4_MOE | 7.03 GB | Tight | 9K | 20K |
Mellum2.1-12B-A2.5B-Thinking on other hardware
- NVIDIA H100Yes
- RTX 5090Yes, just
- RTX 4090Yes, with a GGUF version
- RTX 3090Yes, with a GGUF version
- RTX 5060 Ti 16 GBYes, with a GGUF version
- RTX 3060 12 GBYes, with a GGUF version
- RTX 5060Yes, with a GGUF version
- RTX 4060Yes, with a GGUF version
- Mac with 32 GBYes, with a GGUF version
- Mac with 64 GBYes
- Mac with 128 GBYes