Compare
Four models, one page
Pick the models you are actually choosing between. Every row is scaled within itself, so a bar means something next to its neighbours and nothing across rows.
| Specification | M2.1MiniMax | Claude Opus 5Anthropic |
|---|---|---|
| Released | Dec 23, 2025 | Jul 24, 2026 |
| Context window | 205K | 1M |
| Max output | 131K | 128K |
| Input / 1M | $0.3 | $5 |
| Output / 1M | $1.2 | $25 |
| Cost per task | $0.11 | $2.25 |
| Arena Elo | 1,258 | 1,393 |
| Serving providers | 2 | 5 |
| Parameters | 228.7B | Undisclosed |
| Licence | other | Proprietary |
| Capabilities |
|
|
Metrics side by side
Each row is scaled to the largest value in that row — bars compare within a row, never across rows
- M2.1
- Claude Opus 5
Intelligence Index
M2.132.1Claude Opus 563.1Coding Index
M2.1not measuredClaude Opus 578.0Agentic Index
M2.1not measuredClaude Opus 559.2Output speed
M2.166 t/sClaude Opus 587 t/sContext window
M2.1205KClaude Opus 51MLatency · lower is better
M2.1668msClaude Opus 51.12sBlended price / 1M · lower is better
M2.1$0.525Claude Opus 5$10Cost per task · lower is better
M2.1$0.11Claude Opus 5$2.25
Rows marked “lower is better” still draw a longer bar for a larger number — read the value, not just the length. Arena Elo is in the specification table above instead: it has no meaningful zero, so a bar would flatten the gaps.
View as table
| Metric | M2.1 | Claude Opus 5 |
|---|---|---|
| Intelligence Index | 32.1 | 63.1 |
| Coding Index | — | 78.0 |
| Agentic Index | — | 59.2 |
| Output speed | 66 t/s | 87 t/s |
| Context window | 205K | 1M |
| Latency · lower is better | 668ms | 1.12s |
| Blended price / 1M · lower is better | $0.525 | $10 |
| Cost per task · lower is better | $0.11 | $2.25 |
Evaluation scores
Percentage correct on a common 0–100% scale
- M2.1
- Claude Opus 5
GPQA Diamond
M2.183.0%Claude Opus 593.2%Humanity's Last Exam
M2.123.2%Claude Opus 554.9%SciCode
M2.140.7%Claude Opus 555.7%τ²-bench
M2.185.4%Claude Opus 542.1%Terminal-Bench Hard
M2.128.8%Claude Opus 5not measuredLiveCodeBench
M2.181.0%Claude Opus 5not measuredAA-LCR long context
M2.166.3%Claude Opus 575.7%AIME 2025
M2.182.7%Claude Opus 5not measured
A missing bar means that evaluation was not run for that model — it is not a zero.
View as table
| Evaluation | M2.1 | Claude Opus 5 |
|---|---|---|
| GPQA Diamond | 83.0% | 93.2% |
| Humanity's Last Exam | 23.2% | 54.9% |
| SciCode | 40.7% | 55.7% |
| τ²-bench | 85.4% | 42.1% |
| Terminal-Bench Hard | 28.8% | — |
| LiveCodeBench | 81.0% | — |
| AA-LCR long context | 66.3% | 75.7% |
| AIME 2025 | 82.7% | — |
Estimated from list pricing: 50K input tokens plus 80K output tokens for reasoning models (25K for non-reasoning).