How much VRAM does MiMo V2.6 Flash need?
MiMo V2.6 Flash needs 161 GB for its weights as released (FP4 + FP8), 162 GB with an 8K context and 163 GB with 32K.
- Parameters
- 309B
- Active per token
- 15B
- Longest context
- 1M
- Released
- Sep 2026
- License
- MIT
- Strongest at
- Coding, Agents
Memory at each precision
| Precision | Weights | With 8K | With 32K | With 128K |
|---|---|---|---|---|
| As released (FP4 + FP8) | 161 GB | 162 GB | 163 GB | 165 GB |
| 3-bit | 141 GB | 142 GB | 142 GB | 144 GB |
| 2-bit | 108 GB | 109 GB | 110 GB | 112 GB |
| 1.6-bit | 59.4 GB | 60.4 GB | 60.9 GB | 63.1 GB |
One conversation, 16-bit context cache, runtime overhead included.
What runs it
| Machine | Best fit | Context | Speed | First reply |
|---|---|---|---|---|
| RTX 3060 12 GB | Too big | |||
| RTX 4060 Ti 16 GB | Too big | |||
| RTX 3090 | Too big | |||
| RTX 4090 | Too big | |||
| RTX 5090 | Too big | |||
| 2× RTX 3090 | Too big | |||
| RTX PRO 6000 Blackwell | Too big | |||
| M4 Max 128 GB | Too big | |||
| M5 Max 128 GB | Too big | |||
| Ryzen AI Max+ 395 128 GB | Too big | |||
| DGX Spark 128 GB | 2-bit | 256K | 31 tok/s | 3.2 s |
| M3 Ultra 256 GB | FP4 + FP8 | 928K | 37 tok/s | 8.9 s |
| M5 Ultra 512 GB | FP4 + FP8 | 1M | 49 tok/s | 2.8 s |
| H100 SXM | Too big | |||
| DGX Station GB300 | FP4 + FP8 | 1M | 198 tok/s | 0.3 s |
| 8× H200 NVL | FP4 + FP8 | 1M | 164 tok/s | 0.1 s |
Best fit is the highest quality, up to 8-bit, that leaves 16K of context. First reply is the wait for a 2K prompt. Speeds are rough, plus or minus 25%.
How to run it
- llama.cpp
llama-server -hf ggml-org/MiMo-V2.6-Flash-RL-GGUF- vLLM
vllm serve XiaomiMiMo/MiMo-V2.6-Flash-RL- MLX (Mac)
mlx_lm.chat --model mlx-community/MiMo-V2.6-Flash-RL-mxfp4-q8
Questions
How much VRAM does MiMo V2.6 Flash need?
MiMo V2.6 Flash needs 161 GB for its weights as released (FP4 + FP8), 162 GB with an 8K context and 163 GB with 32K.
Can MiMo V2.6 Flash run on an RTX 4090?
Not on one card. The smallest machine here that runs it is the DGX Spark 128 GB.