Model reference · open weights
M1 is an open-weight language model from togethercomputer. M1-3B (BF16) weighs 6.9 GB; the smallest configuration that runs it is RTX 3060 12 GB.
Summary of the togethercomputer/M1-3B model card, 2026-10-01
What it is
| Released by | togethercomputer |
|---|---|
| Released | 2025-05-02 |
| Parameters | 3.4B |
| VRAM | 6.9 GB for the weights |
What it runs on
| Card | Requests at once | Context max | Memory | |
|---|---|---|---|---|
| 8K each | 32K each | |||
| RTX 3060 12 GB | 4 | 1 | 34K | 11.6 GB |
| RTX 4060 Ti 16 GB | 8 | 2 | 66K | 15.4 GB |
| RTX 3090 24 GB | 16 | 4 | all 128K | 23.4 GB |
| RTX 4090 24 GB | 16 | 4 | all 128K | 23.4 GB |
| RTX 5090 32 GB | 24 | 6 | all 128K | 31.0 GB |
| L40S 48 GB | 38 | 9 | all 128K | 44.0 GB |
| A100 80 GB | 75 | 18 | all 128K | 78.2 GB |
| H100 80 GB | 70 | 17 | all 128K | 78.1 GB |
| RTX PRO 6000 Blackwell 96 GB | 87 | 21 | all 128K | 93.8 GB |
| DGX Spark (GB10) 128 GB unified | 102 | 25 | all 128K | 107 GB |
| H200 141 GB | 134 | 33 | all 128K | 138 GB |
| B200 180 GB | 174 | 43 | all 128K | 176 GB |
| 2× RTX 3060 12 GB tensor parallel | 15 | 3 | 127K | 11.6 GB a card |
| Requests at once | 8K tokens each | 32K tokens each |
|---|---|---|
| 1 | 8.5 GB | 11.3 GB |
| 5 | 12.3 GB | 26.4 GB |
| 8 | 15.1 GB | 37.6 GB |
| 16 | 22.6 GB | 67.7 GB |
| 32 | 37.6 GB | 128 GB |
| 64 | 67.7 GB | 248 GB |
One card, with vLLM's small-card settings.
From the model card
This is the model is trained using paper, M1: Towards Scalable Test-Time Compute with Mamba Reasoning Models.
| Model | AIME 2025 | AIME 2024 | MATH 500 | AMC 2023 | OlympiadBench |
|---|---|---|---|---|---|
| Qwen2.5-Math-7B-Instruct (Transformer) | – | 13.3 | 79.8 | 50.6 | 40.7 |
| rStar-Math-7B (Transformer) | – | 26.7 | 78.4 | 47.5 | 47.1 |
| Eurus-2-7B-PRIME (Transformer) | – | 26.7 | 79.2 | 57.8 | 42.1 |
| Qwen2.5-7B-SimpleRL (Transformer) | – | 26.7 | 82.4 | 62.5 | 43.3 |
| DeepSeek-R1-Distill-Qwen-1.5B (Transformer) | 23.0 | 28.8 | 82.8 | 62.9 | 43.3 |
| M1-3B (Mamba Hybrid Models) | 23.5 | 28.5 | 84.0 | 62.8 | 47.3 |
Code: https://github.com/jxiw/M1
@article{wang2025m1scalabletesttimecompute,
title={M1: Towards Scalable Test-Time Compute with Mamba Reasoning Models},
author={Junxiong Wang and Wen-Ding Li and Daniele Paliotta and Daniel Ritter and Alexander M. Rush and Tri Dao},
journal={arXiv preprint arXiv:2504.10449},
year={2025},
url={https://arxiv.org/abs/2504.10449},
}Quoted from the model card on Hugging Face. The full card is behind the Hugging Face link above.