Model comparison
MiniMax-M3 vs o1
MiniMax-M3 is the stronger model overall, scoring 43.8 to 40.9 on the Noometry Index.
Last verified . 25 shared benchmarks.
MiniMaxSummary
- They share 25 benchmarks with published results for both. MiniMax-M3 scores higher in 7 categories and o1 in 3 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where MiniMax-M3 leads 58.4 to 41.5.
- The biggest single-benchmark swing is GPQA Diamond: 90.9% for MiniMax-M3 and 76.8% for o1.
- MiniMax-M3 is cheaper at $0.30 / $1.20 per million input/output tokens, against $15 / $60 for o1.
- MiniMax-M3 accepts more context: 1M tokens versus 200K.
- MiniMax-M3 has downloadable open weights; the other is API-only.
Side by side
| MiniMax-M3 | o1 | |
|---|---|---|
| Provider | MiniMax | OpenAI |
| Noometry Index | 43.8 | 40.9 |
| Released | 2026-06-01 | 2024-09-12 |
| Weights | Open | Proprietary |
| Context window | 1M | 200K |
| Max output | 512K | 100K |
| Input $ / M tokens | $0.30 | $15 |
| Output $ / M tokens | $1.20 | $60 |
| Results tracked | 41 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o1 leads
MiniMax-M3: 41.8 (#118), o1: 46.1 (#70)
| Benchmark | MiniMax-M3 | o1 |
|---|---|---|
| LMArena Coding | 1469 | 1367 |
| FrontierCode | 14.7% | — |
| Aider Polyglot | — | 61.7% |
| LMArena WebDev | 1482 | — |
| SciCode | 47.1% | — |
| WeirdML | — | 47.6% |
| LiveBench Coding | — | 69.7% |
| CadEval | — | 56% |
| ALE-Bench | 640.02 | — |
| HumanEval+ | — | 89% |
| MBPP+ | — | 80.2% |
Agentic & Tool Use o1 leads
MiniMax-M3: 22.6 (#130), o1: 24.6 (#117)
| Benchmark | MiniMax-M3 | o1 |
|---|---|---|
| APEX-Agents | 37.7% | — |
| OSWorld 2.0 | 4.6% | — |
| Cybench | — | 10% |
| GBAEval | 0.9% | — |
| METR Time Horizons | — | 51.1% |
| Vending-Bench 2 | 2,158 | — |
Reasoning MiniMax-M3 leads
MiniMax-M3: 30.1 (#87), o1: 27.9 (#111)
| Benchmark | MiniMax-M3 | o1 |
|---|---|---|
| SimpleBench | 45.8% | 41.7% |
| Chess Puzzles | 14% | 15% |
| LMArena Hard Prompts | 1447 | 1371 |
| DTBench | 78.9% | 74.7% |
| LMCA | 33.7% | 22.3% |
| Epoch Capabilities Index | 146.95 | 141.91 |
| NYT Connections (extended) | 65.1% | — |
| ARC-AGI-1 | — | 30.7% |
| CritPt | 3.7% | — |
| EnigmaEval | — | 5.7% |
| LiveBench Reasoning | — | 91.6% |
| Mystery Game Puzzles | 8% | — |
| LiveBench Data Analysis | — | 65.5% |
| Surface Evolver Bench | 55% | — |
| ForecastBench | 61.4 | — |
| LiveBench | — | 75.7% |
Math MiniMax-M3 leads
MiniMax-M3: 40.0 (#95), o1: 36.1 (#175)
| Benchmark | MiniMax-M3 | o1 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 71.1% | 73.3% |
| LMArena Math | 1429 | 1388 |
| FrontierMath (Tiers 1-3) | — | 14.7% |
| ProofBench | 18% | — |
| LiveBench Math | — | 80.3% |
| MATH Level 5 | — | 94.7% |
| FrontierMath (Feb 2025 set) | — | 9.3% |
Knowledge MiniMax-M3 leads
MiniMax-M3: 58.4 (#35), o1: 41.5 (#110)
| Benchmark | MiniMax-M3 | o1 |
|---|---|---|
| GPQA Diamond | 90.9% | 76.8% |
| LMArena Expert | 1461 | 1361 |
| Humanity's Last Exam | — | 8% |
| SimpleQA Verified | — | 41.1% |
| Confabulations | — | 11.7% |
Multimodal MiniMax-M3 leads
MiniMax-M3: 40.2 (#51), o1: 34.2 (#93)
| Benchmark | MiniMax-M3 | o1 |
|---|---|---|
| LMArena Vision | 1253 | 1168 |
| GeoBench | — | 80% |
| VPCT | — | 37% |
| LMArena Document | 1435 | — |
| SpatialViz-Bench | — | 41.4% |
Multilingual MiniMax-M3 leads
MiniMax-M3: 53.0 (#75), o1: 48.6 (#142)
| Benchmark | MiniMax-M3 | o1 |
|---|---|---|
| LMArena Non-English | 1420 | 1358 |
| LMArena Chinese | 1463 | 1394 |
| LMArena French | 1447 | 1344 |
| LMArena German | 1426 | 1337 |
| LMArena Japanese | 1381 | 1346 |
| LMArena Korean | 1372 | 1396 |
| LMArena Russian | 1428 | 1356 |
| LMArena Spanish | 1432 | 1345 |
Instruction Following Too close to call
MiniMax-M3: 75.5 (#62), o1: 74.8 (#86)
| Benchmark | MiniMax-M3 | o1 |
|---|---|---|
| LMArena Instruction Following | 1433 | 1367 |
| LiveBench Instruction Following | — | 81.5% |
Long Context o1 leads
MiniMax-M3: 44.2 (#72), o1: 50.3 (#9)
| Benchmark | MiniMax-M3 | o1 |
|---|---|---|
| LMArena Longer Query | 1445 | 1378 |
| Fiction.LiveBench | — | 83.3% |
Writing & Preference MiniMax-M3 leads
MiniMax-M3: 62.1 (#83), o1: 55.6 (#144)
| Benchmark | MiniMax-M3 | o1 |
|---|---|---|
| LMArena Text | 1433 | 1366 |
| LMArena Creative Writing | 1404 | 1348 |
| LMArena Multi-Turn | 1442 | 1369 |
| Short-Story Creative Writing | — | 70.2% |
| EQ-Bench 4 | 1150 | — |
| LiveBench Language | — | 65.4% |
Frequently asked questions
Is MiniMax-M3 better than o1?
MiniMax-M3 is the stronger model overall, scoring 43.8 to 40.9 on the Noometry Index.
Which is cheaper, MiniMax-M3 or o1?
MiniMax-M3 is cheaper. It lists at $0.30 per million input tokens and $1.20 per million output tokens; o1 lists at $15 and $60.
Is MiniMax-M3 or o1 better for coding?
o1 scores higher on coding benchmarks: 46.1 versus 41.8 in the Noometry coding category.
Which has the bigger context window?
MiniMax-M3 does, with 1M tokens against 200K.
How many benchmarks do MiniMax-M3 and o1 share?
25 benchmarks have published results for both models. MiniMax-M3 has 41 scored results on Noometry and o1 has 52.