MiniMax-M2.7: benchmark scores, price and ranking
MiniMax-M2.7 from MiniMax ranks #568 of 743 LLMs on the LLMs Tiger aggregate benchmark score (45.6), based on 10 benchmark results captured up to 2026-10-13.
- Organisation
- MiniMax
- Type
- Open-weight / local model
- Open weights
- Yes
- Released
- 2026-03-18 (source: epoch)
- Cheapest API price
- $0.250 in / $1.00 out per million tokens (via deepinfra, observed 2026-10-13T19:47Z)
- Aggregate rank
- #568 of 743 (score 45.6)
Relative to other models, MiniMax-M2.7 is strongest on Epoch ECI (#26 of 204) and weakest on ProofBench (#43 of 46).
Across the 8 benchmarks where at least ten models are scored, MiniMax-M2.7 is in the top 10% on 0 and in the top half on 3.
Its cheapest listed input price, $0.250 per million tokens, is lower than that of 64% of the 427 models with a listed price.
MiniMax-M2.7 benchmark results
| Benchmark | Score | Rank | Source | Captured |
|---|---|---|---|---|
| Epoch ECI | 146 | #26 of 204 | Epoch AI Benchmarking Hub | 2026-10-13 |
| Terminal-Bench 2.0 | 45.1% | #9 of 28 | Epoch AI Benchmarking Hub | 2026-10-13 |
| SciCode | 47.0% | #46 of 106 | Epoch AI Benchmarking Hub | 2026-10-13 |
| WebDev Arena | 1,398 | #37 of 67 | Epoch AI Benchmarking Hub | 2026-10-13 |
| CritPt | 0.6% | #74 of 115 | Epoch AI Benchmarking Hub | 2026-10-13 |
| WeirdML | 37.0% | #74 of 109 | Epoch AI Benchmarking Hub | 2026-10-13 |
| ALE-Bench | 599 | #54 of 79 | Epoch AI Benchmarking Hub | 2026-10-13 |
| ProofBench | 3.0% | #43 of 46 | Epoch AI Benchmarking Hub | 2026-10-13 |
| GBA-Eval | 0.0% | #7 of 7 | Epoch AI Benchmarking Hub | 2026-10-13 |
| ExploitBench | 13.3% | #4 of 4 | Epoch AI Benchmarking Hub | 2026-10-13 |
About MiniMax-M2.7
MiniMax-M2.7 is an open-weight model that can run on your own hardware from MiniMax. It was released on 2026-03-18, according to Epoch AI. The cheapest API listing we track charges $0.250 per million input tokens and $1.00 per million output tokens, through deepinfra. We hold 10 benchmark results for it across 4 areas, and none is in the top tenth of its board.
Where MiniMax-M2.7 stands
On our aggregate score, MiniMax-M2.7 is #568 of 743 ranked models, in the lower half. Its score is 45.6; the scale is centred on 50, so it sits 4.4 points below the typical model. That beats 24% of the ranked field. Among the 6 ranked MiniMax models it is #6. The next one up is MiniMax-M3 at 45.6.
MiniMax-M2.7, benchmark by benchmark
Epoch ECI puts MiniMax-M2.7 at 146, #26 of 204. That beats 87% of the models tested, but 11 points short of Qwen 3.8 Max (156). The results either side are Kimi K2 Thinking (146) and GLM-5 (146). It covers Epoch Capabilities Index: Epoch's own aggregate of ~40 benchmarks. The board median is 127, so MiniMax-M2.7 is 19 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
On Terminal-Bench 2.0, MiniMax-M2.7 reaches 45.1%. It ranks #9 of 28, ahead of 68% of the field, and 24.4 points behind Gemini 3 Pro Preview at 69.4%. The results either side are GLM-5 (52.4%) and Kimi K2.5 (43.2%). It covers real terminal tasks (agent + model). The board median is 37.3%, so MiniMax-M2.7 is 7.8 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
MiniMax-M2.7 manages 47.0% on SciCode, good for #46 of 106. Claude Opus 5.5 (max) leads at 66.9%, 19.9 points further up, and 57% of models sit below. The results either side are Inkling (xhigh) (47.0%) and GPT-5.4 nano (xhigh) (46.9%). It covers scientific research code. The board median is 44.3%, so MiniMax-M2.7 is 2.7 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
MiniMax-M2.7 lands below the middle on WebDev Arena: 1,398, #37 of 67, with 54% of the field scoring higher. Claude Opus 5.5 (max) leads at 1,820; the closest result above is Qwen3.5 Plus at 1,399. It covers building web apps, crowd-voted (Elo). The board median is 1,433, so MiniMax-M2.7 is 36 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
MiniMax-M2.7 lands below the middle on CritPt: 0.6%, #74 of 115, with 63% of the field scoring higher. GPT-5.6 Sol (max) leads at 32.3%; the closest result above is o4-mini (high) at 0.6%. It covers unpublished research-level physics problems. The board median is 2.9%, so MiniMax-M2.7 is 2.3 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
MiniMax-M2.7 lands below the middle on WeirdML: 37.0%, #74 of 109, with 67% of the field scoring higher. GPT-6 Astra (max) leads at 93.3%; the closest result above is grok-3 at 37.2%. It covers unusual machine-learning tasks. The board median is 41.2%, so MiniMax-M2.7 is 4.2 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
MiniMax-M2.7 lands below the middle on ALE-Bench: 599, #54 of 79, with 67% of the field scoring higher. GPT-6 Astra (max) leads at 2,951; the closest result above is MiniMax-M2.5 at 618. It covers AtCoder heuristic-optimisation contests (performance rating). The board median is 764, so MiniMax-M2.7 is 165 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
ProofBench is a weak spot. MiniMax-M2.7 scores 3.0%, #43 of 46, behind 91% of the models measured. The leader, Claude Fable 5.1 (max), reaches 100.0%. It covers writing mathematical proofs. The board median is 21.0%, so MiniMax-M2.7 is 18.0 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
The full table above also lists the other 2 results, including boards with fewer than ten models, where a rank says little.
Strengths and weaknesses
Software engineering: solid, ahead of 50% of the field on average across 4 benchmarks. Its best board here is Terminal-Bench 2.0 (#9 of 28) and its weakest is ALE-Bench (#54 of 79). Reasoning, maths and science: weak, ahead of 21% of the field on average across 2 benchmarks. Its best board here is CritPt (#74 of 115) and its weakest is ProofBench (#43 of 46). General knowledge and chat: solid, ahead of 60% of the field on average across 2 benchmarks. Its best board here is Epoch ECI (#26 of 204) and its weakest is WeirdML (#74 of 109).
Price and value
The median input price across the 427 priced models is $0.500 per million tokens, so MiniMax-M2.7 costs less than the typical listing. Models that score within 1.5 points or better and cost less: Qwen2.5-Coder-7B-Instruct ($0.010, 49.5), DeepSeek V4.1 Flash (high) ($0.010, 48.9), Qwen2.5-Coder-3B-Instruct ($0.010, 46.5). Similar scores cost more on DeepSeek-V3.2 (Thinking; Novita) ($0.259) and MiniMax-M2.5 ($0.270).
Release history
The MiniMax model released before it was MiniMax-M2.5 on 2026-02-12, which scores 46.8. The next one was MiniMax-M3 on 2026-06-01, at 45.6.
Reading these numbers
The 10 scores on this page come from 1 source: Epoch AI Benchmarking Hub (10). No official leaderboard supplied any of them, so treat them as third-party measurements.
Models ranked nearby
- #565 Llama 4 Scout Instruct (45.6)
- #566 MiniMax-M3 (45.6)
- #567 open-mistral-nemo-2407 (45.6)
- #569 ollama/mistral-nemo:12b-instruct-2407-q4_K_M (45.6)
- #570 Llama 4 Maverick (FP8) (45.6)
- #571 mistral-large-2411 (45.6)
More from MiniMax
- MiniMax-M1-80k (#160, 50.9)
- MiniMax-M2.1 (#228, 49.9)
- MiniMax-M2 (#330, 48.7)
- MiniMax-M2.5 (#484, 46.8)
- MiniMax-M3 (#566, 45.6)
Compare it on the live leaderboard → · How the aggregate score works