MiniMax model product
MiniMax-M1 is an open-source, large-scale reasoning model that uses a hybrid-attention architecture for efficient long-context processing. It supports up to a 1 million token context window and 80,000-token reasoning output, matching Gemini 2.5 Pro’s scale while being highly cost-effective. Its Lightning Attention mechanism reduces compute requirements to about 30% of DeepSeek R1’s, and a new reinforcement learning algorithm, CISPO, doubles convergence speed compared to other RL methods. Trained on 512 H800s over three weeks, M1 achieves near state-of-the-art results across software engineering, long-context, and tool-use benchmarks, outperforming most open models and rivaling top closed systems.
Updated Aug 10, 2026. Default version: MiniMax M1 80K
Structured fields from the published default version.
This profile uses the latest version under this unique model that has a calculated LLMBoard score. Arena and price are excluded.
Published benchmark records for the scored version MiniMax M1 80K.
| OpenAI-MRCR: 2 needle 1M | 0.6 | 2 | 5 | 75.0% | C | |
| TAU-bench Airline | 0.6 | 2 | 23 | 95.5% | C | |
| OpenAI-MRCR: 2 needle 128k | 0.7 | 3 | 9 | 75.0% | C | |
| LongBench v2 | 0.6 | 5 | 17 | 75.0% | C | |
| ZebraLogic | 0.9 | 7 | 8 | 14.3% | C | |
| MATH-500 | 1.0 | 11 | 32 | 67.7% | C | |
| AIME 2024 | 0.9 | 15 | 53 | 73.1% | C | |
| TAU-bench Retail | 0.6 | 18 | 25 | 29.2% | C | |
| Multi-Challenge | 0.4 | 19 | 29 | 35.7% | C | |
| LiveCodeBench | 0.7 | 27 | 73 | 63.9% | C | |
| SimpleQA | 0.2 | 35 | 46 | 24.4% | C | |
| MMLU-Pro | 0.8 | 47 | 129 | 64.1% | C | |
| AIME 2025 | 0.8 | 77 | 114 | 32.7% | C | |
| SWE-Bench Verified | 0.6 | 82 | 104 | 21.4% | C | |
| Humanity's Last Exam | 0.1 | 84 | 92 | 8.8% | C | |
| GPQA | 0.7 | 128 | 233 | 45.3% | C |
Preference and agent-evaluation signals from published Arena datasets.
The default version has no published Arena rows, or its source alias has not been resolved.
Official vendor API PAYG pricing is summarized first. The table then lists individual provider offerings without treating their minimum as the official price.
| NanoGPT | MiniMaxAI/MiniMax-M1-80k | global | $0.6052 | $2.42 | 1M |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
All published versions linked to this unique model. The score columns identify the version used by the current overall ranking.
| MiniMax M1 80K | 56.3 | 456B | 1M | 40K | No | MIT |
A concise description based on the published model registry.
MiniMax-M1 is an open-source, large-scale reasoning model that uses a hybrid-attention architecture for efficient long-context processing. It supports up to a 1 million token context window and 80,000-token reasoning output, matching Gemini 2.5 Pro’s scale while being highly cost-effective. Its Lightning Attention mechanism reduces compute requirements to about 30% of DeepSeek R1’s, and a new reinforcement learning algorithm, CISPO, doubles convergence speed compared to other RL methods. Trained on 512 H800s over three weeks, M1 achieves near state-of-the-art results across software engineering, long-context, and tool-use benchmarks, outperforming most open models and rivaling top closed systems.
Use the benchmark, Arena and pricing sections above as separate evidence. A missing field means the current data snapshot does not support that claim.
Data snapshot: 2026-08-07. Editorial model content is not available in the backend.
Open a comparison with the three ranked models immediately above and below this model.
Recommendations prioritize the same model type and family, then the closest published LLMBoard score.
Common questions about MiniMax M1 80K.
MiniMax M1 80K's default version was released on Jun 16, 2025.
No official standard PAYG price is currently available for MiniMax M1 80K. The lowest tracked third-party offer starts at $0.6052 input and $2.42 output via NanoGPT.
MiniMax M1 80K is published under MiniMax in the model registry.
The default version has a 1M token context window.
No. The default version is not marked as having publicly available weights.
1 published provider offerings are linked to the default version.
Nearby ranked alternatives include Qwen2.5 32B, GPT-OSS-120B, GPT-5.4-mini.