MiniMax M2.5vsMiniMax M2
Across 9 shared benchmarks, MiniMax M2.5 leads overall: MiniMax M2.5 wins 8, MiniMax M2 wins 1, with 0 ties and an average score difference of +9.41.
MiniMax M2.5
MiniMaxAI · 2026-02-12 · Reasoning model
MiniMax M2
MiniMaxAI · 2025-10-27 · Chat model
MiniMax M2.58 wins(89%)(11%)1 winMiniMax M2
Benchmark scores
Grouped by capability, sorted by largest gap within each. 9 shared benchmarks.
Agent Level Benchmark
MiniMax M2.5 2/2| Benchmark | MiniMax M2.5 | MiniMax M2 | Diff |
|---|---|---|---|
| τ²-Bench - Telecom | 97.8012 / 264Thinking (With Tools) | 8777 / 264Thinking (With Tools) | +10.80 |
| Terminal Bench Hard | 34.8075 / 244Thinking (With Tools) | 25.80126 / 244Thinking (With Tools) | +9 |
General Knowledge
MiniMax M2.5 2/2| Benchmark | MiniMax M2.5 | MiniMax M2 | Diff |
|---|---|---|---|
| HLE | 20.50281 / 563Thinking (No Tools) · Text only | 13.70351 / 563Thinking (No Tools) · Text only | +6.80 |
| CritPt | 1.10147 / 200Thinking (No Tools) | 0.90158 / 200Thinking (No Tools) | +0.20 |
AI Agent - Information Search
MiniMax M2.5 1/1| Benchmark | MiniMax M2.5 | MiniMax M2 | Diff |
|---|---|---|---|
| BrowseComp | 76.3025 / 57Thinking (With Tools) | 4449 / 57Thinking (With Tools) | +32.30 |
Coding and Software Engineer
MiniMax M2.5 1/1| Benchmark | MiniMax M2.5 | MiniMax M2 | Diff |
|---|---|---|---|
| SWE-bench Verified | 80.2014 / 116Thinking (With Tools) | 69.4064 / 116Thinking (With Tools) | +10.80 |
General Evaluation
MiniMax M2.5 1/1| Benchmark | MiniMax M2.5 | MiniMax M2 | Diff |
|---|---|---|---|
| GPQA Diamond | 85.20157 / 462Thinking (No Tools) | 78256 / 462Thinking (No Tools) | +7.20 |
Instruction Following
MiniMax M2 1/1| Benchmark | MiniMax M2.5 | MiniMax M2 | Diff |
|---|---|---|---|
| IF Bench | 71.6055 / 282Thinking (No Tools) | 72.3052 / 282Thinking (No Tools) | -0.70 |
Math and Reasoning
MiniMax M2.5 1/1| Benchmark | MiniMax M2.5 | MiniMax M2 | Diff |
|---|---|---|---|
| AIME2025 | 86.3073 / 215Thinking (No Tools) | 78101 / 215Thinking (No Tools) | +8.30 |
Specs
| Field | MiniMax M2.5 | MiniMax M2 |
|---|---|---|
| Publisher | MiniMaxAI | MiniMaxAI |
| Release date | 2026-02-12 | 2025-10-27 |
| Model type | Reasoning model | Chat model |
| Architecture | MoE | MoE |
| Parameters | 229B | 230B |
| Context length | 128K | 205K |
| Max output | Not available | Not available |
API pricing
Prices use DataLearner records when available; missing fields are not inferred.
| Item | MiniMax M2.5 | MiniMax M2 |
|---|---|---|
| Text input | $0.3 / 1M tokens | ¥2.1 / 1M tokens |
| Text output | $2.4 / 1M tokens | ¥8.4 / 1M tokens |
| Cache read | Not public | ¥0.21 / 1M tokens |
| Cache write | Not public | ¥2.625 / 1M tokens |
Summary
- MiniMax M2.5leads in:Agent Level Benchmark (2/2), General Knowledge (2/2), AI Agent - Information Search (1/1), Coding and Software Engineer (1/1), General Evaluation (1/1), Math and Reasoning (1/1)
- MiniMax M2leads in:Instruction Following (1/1)
On average across the 9 shared benchmarks, MiniMax M2.5 scores 9.41 higher.
Largest single-benchmark gap: BrowseComp — MiniMax M2.5 76.30 vs MiniMax M2 44 (+32.30).
Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.