MiniMax, open weights
MiniMax-M2
MiniMax-M2 by MiniMax ranks 204th of 354 ranked models on the Noometry Index as of October 2026, with a score of 37.4. Its strongest category is agentic & tool use, where it ranks 109th. API pricing starts at $0.30 per million input tokens and $1.20 per million output tokens, with a 205K-token context window.
Last verified
Specifications
- Noometry rank
- #204 of 354
- Index score
- 37.4
- Evidence
- Confirmed 21 results
- Provider
MiniMax- Released
- October 27, 2025
- Weights
- Open weights
- Reasoning
- Yes
- Context window
- 205K
- Max output
- 131K
- Input price
- $0.30 / M
- Output price
- $1.20 / M
- Blended price
- $0.52 / M
- Output speed
- 17 tokens/s Kagi
- Value
- #84 of 219
- Knowledge cutoff
- Unknown
- Input
- text
- Hugging Face
- MiniMaxAI/MiniMax-M2
Category scores
Each category score combines every public result we have in that category.
- Coding 39.3
- Agentic & Tool Use 25.1
- Reasoning 19.4
- Math 37.3
- Knowledge 37.0
- Multilingual 45.3
- Instruction Following 70.2
- Long Context 40.5
- Writing & Preference 53.0
| Category | Score | Rank | Results |
|---|---|---|---|
| Coding | 39.3 | #159 | 3 |
| Agentic & Tool Use | 25.1 | #109 | 1 |
| Reasoning | 19.4 | #258 | 3 |
| Math | 37.3 | #160 | 1 |
| Knowledge | 37.0 | #163 | 1 |
| Multilingual | 45.3 | #171 | 1 |
| Instruction Following | 70.2 | #166 | 1 |
| Long Context | 40.5 | #153 | 1 |
| Writing & Preference | 53.0 | #162 | 3 |
Strengths and weaknesses
Categories where MiniMax-M2 places highest and lowest among the models ranked in each, with its score against that category's median.
Strongest categories
| Category | Score | vs median | Rank |
|---|---|---|---|
| Coding | 39.3 | +0.5 | #159 of 340, top 47% |
| Math | 37.3 | +0.8 | #160 of 327, top 49% |
| Long Context | 40.5 | −0.4 | #153 of 296, top 52% |
Weakest categories
| Category | Score | vs median | Rank |
|---|---|---|---|
| Reasoning | 19.4 | −4.2 | #258 of 350, top 74% |
| Agentic & Tool Use | 25.1 | −5.2 | #109 of 154, top 71% |
| Multilingual | 45.3 | −2.1 | #171 of 297, top 58% |
Closest competitors
The models ranked just above and below MiniMax-M2. When scores are this close, price and speed are often the better way to choose.
| Model | Rank | Score | Blended $/M | Speed | |
|---|---|---|---|---|---|
| DeepSeek-V2.5 (Sep 2024) | #200 | 37.6 | — | — | Compare |
| Qwen3.6 35B-A3B | #201 | 37.6 | $0.56 | — | Compare |
| Hunyuan Large Vision | #202 | 37.6 | — | — | Compare |
| Llama 3.1 Nemotron 70b Instruct | #203 | 37.6 | — | — | Compare |
| Granite 4.1 8b | #205 | 37.4 | — | — | Compare |
| Gemma 3n E4b IT | #206 | 37.3 | — | 13 | Compare |
| Step 3.7 Flash | #207 | 37.3 | $0.42 | — | Compare |
| GPT-4.5 | #208 | 37.2 | — | — | Compare |
Sponsored placements are available on pages like this one. Advertise on Noometry
Benchmark results
Every published result we track, with its source. Bold rows are the ones used for ranking; where several exist we prefer independent runs over self-reported numbers.
Coding
| Benchmark | Score | Position | Setting | Source | Date |
|---|---|---|---|---|---|
| SWE-bench Verified (bash only) | 61% | #19 of 39, top 49% | SWE-bench | 2025-11-24 | |
| LMArena WebDev | 1297 | #97 of 113, top 86% | LMArena | 2026-10-08 | |
| LMArena Coding | 1370 | #157 of 294, top 54% | LMArena | 2026-10-08 |
Agentic & Tool Use
| Benchmark | Score | Position | Setting | Source | Date |
|---|---|---|---|---|---|
| Terminal-Bench | 30% | #32 of 41, top 79% | Epoch AI | ||
| Vending-Bench 2 | 160.6 | #53 of 60, top 89% | Epoch AI |
Reasoning
| Benchmark | Score | Position | Setting | Source | Date |
|---|---|---|---|---|---|
| Kagi LLM Benchmark | 57.8% | #46 of 99, top 47% | Kagi LLM Benchmark | ||
| NYT Connections (extended) | 14.8% | #83 of 91, top 92% | Lech Mazur benchmarks | ||
| LMArena Hard Prompts | 1357 | #155 of 297, top 53% | LMArena | 2026-10-08 |
Math
| Benchmark | Score | Position | Setting | Source | Date |
|---|---|---|---|---|---|
| LMArena Math | 1352 | #164 of 285, top 58% | LMArena | 2026-10-08 |
Knowledge
| Benchmark | Score | Position | Setting | Source | Date |
|---|---|---|---|---|---|
| LMArena Expert | 1337 | #162 of 273, top 60% | LMArena | 2026-10-08 |
Multilingual
| Benchmark | Score | Position | Setting | Source | Date |
|---|---|---|---|---|---|
| LMArena Non-English | 1313 | #171 of 297, top 58% | LMArena | 2026-10-08 | |
| LMArena Chinese | 1366 | #157 of 285, top 56% | LMArena | 2026-10-08 | |
| LMArena French | 1335 | #149 of 223, top 67% | LMArena | 2026-10-08 | |
| LMArena German | 1355 | #126 of 231, top 55% | LMArena | 2026-10-08 | |
| LMArena Russian | 1331 | #159 of 283, top 57% | LMArena | 2026-10-08 | |
| LMArena Spanish | 1326 | #153 of 226, top 68% | LMArena | 2026-10-08 |
Instruction Following
| Benchmark | Score | Position | Setting | Source | Date |
|---|---|---|---|---|---|
| LMArena Instruction Following | 1328 | #159 of 298, top 54% | LMArena | 2026-10-08 |
Long Context
| Benchmark | Score | Position | Setting | Source | Date |
|---|---|---|---|---|---|
| LMArena Longer Query | 1331 | #163 of 291, top 57% | LMArena | 2026-10-08 |
Writing & Preference
| Benchmark | Score | Position | Setting | Source | Date |
|---|---|---|---|---|---|
| LMArena Text | 1340 | #165 of 297, top 56% | LMArena | 2026-10-08 | |
| LMArena Creative Writing | 1286 | #180 of 295, top 62% | LMArena | 2026-10-08 | |
| LMArena Multi-Turn | 1361 | #148 of 295, top 51% | LMArena | 2026-10-08 |
API pricing by provider
| Route | Input $/M | Output $/M | Cached input $/M | Checked |
|---|---|---|---|---|
| bedrock | $0.30 | $1.20 | — | 2026-10-10 |
| minimax | $0.30 | $1.20 | — | 2026-10-10 |
| openrouter | $0.30 | $1.20 | — | 2026-10-10 |
Compare MiniMax-M2
- MiniMax-M2 vs MiniMax M1
- MiniMax-M2 vs Llama 3.1 Nemotron 70b Instruct
- MiniMax-M2 vs Granite 4.1 8b
- MiniMax-M2 vs Hunyuan Large Vision
- MiniMax-M2 vs Gemma 3n E4b IT
- MiniMax-M2 vs Qwen3.6 35B-A3B
- MiniMax-M2 vs Step 3.7 Flash
- MiniMax-M2 vs GPT-6 Astra
- MiniMax-M2 vs Claude Fable 5.1
- MiniMax-M2 vs Gemini 3.8 Flash
- MiniMax-M2 vs Kimi K3
- MiniMax-M2 vs Grok 4.6
- MiniMax-M2 vs Qwen3.8 Max
- MiniMax-M2 vs GLM-5.3
Other MiniMax models
Frequently asked questions
How good is MiniMax-M2?
MiniMax-M2 by MiniMax ranks 204th of 354 ranked models on the Noometry Index as of October 2026, with a score of 37.4. Its strongest category is agentic & tool use, where it ranks 109th. API pricing starts at $0.30 per million input tokens and $1.20 per million output tokens, with a 205K-token context window.
How much does MiniMax-M2 cost?
MiniMax-M2 costs $0.30 per million input tokens and $1.20 per million output tokens on MiniMax's own API.
What is MiniMax-M2's context window?
MiniMax-M2 accepts up to 205K tokens of input and can write up to 131K tokens in one response.
Is MiniMax-M2 open source?
Yes. MiniMax-M2's weights are downloadable from Hugging Face (MiniMaxAI/MiniMax-M2); check the license for commercial terms.
How fast is MiniMax-M2?
MiniMax-M2 generated about 17 output tokens per second in the Kagi LLM Benchmark's timed runs. Speed varies by provider, load and reasoning effort.
What are MiniMax-M2's strengths and weaknesses?
Relative to other ranked models, MiniMax-M2 places best in coding, math, long context and lowest in reasoning, agentic & tool use, multilingual.
What is MiniMax-M2 best at?
Its best category is agentic & tool use, where it ranks 109th on Noometry.