Model comparison
MiniMax-M3 vs o4-mini
MiniMax-M3 is the stronger model overall, scoring 43.8 to 41.6 on the Noometry Index.
Last verified . 29 shared benchmarks.
MiniMaxSummary
- They share 29 benchmarks with published results for both. MiniMax-M3 scores higher in 6 categories and o4-mini in 4 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where MiniMax-M3 leads 58.4 to 43.6.
- The biggest single-benchmark swing is Chess Puzzles: 14% for MiniMax-M3 and 26% for o4-mini.
- MiniMax-M3 is cheaper at $0.30 / $1.20 per million input/output tokens, against $1.10 / $4.40 for o4-mini.
- 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 | o4-mini | |
|---|---|---|
| Provider | MiniMax | OpenAI |
| Noometry Index | 43.8 | 41.6 |
| Released | 2026-06-01 | 2025-04-16 |
| Weights | Open | Proprietary |
| Context window | 1M | 200K |
| Max output | 512K | 100K |
| Input $ / M tokens | $0.30 | $1.10 |
| Output $ / M tokens | $1.20 | $4.40 |
| Results tracked | 41 | 60 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
MiniMax-M3: 41.8 (#118), o4-mini: 40.9 (#127)
| Benchmark | MiniMax-M3 | o4-mini |
|---|---|---|
| LMArena Coding | 1469 | 1368 |
| ALE-Bench | 640.02 | 826.17 |
| FrontierCode | 14.7% | — |
| SWE-bench Verified (bash only) | — | 45% |
| Aider Polyglot | — | 72% |
| LMArena WebDev | 1482 | — |
| SciCode | 47.1% | — |
| GSO | — | 3.6% |
| WeirdML | — | 52.6% |
| CadEval | — | 62% |
| AlgoTune | — | 1.72 |
Agentic & Tool Use o4-mini leads
MiniMax-M3: 22.6 (#130), o4-mini: 32.6 (#61)
| Benchmark | MiniMax-M3 | o4-mini |
|---|---|---|
| APEX-Agents | 37.7% | — |
| Berkeley Function Calling Leaderboard | — | 53.2% |
| OSWorld 2.0 | 4.6% | — |
| GDPval | — | 25.3% |
| GBAEval | 0.9% | — |
| METR Time Horizons | — | 63.9% |
| Vending-Bench 2 | 2,158 | — |
Reasoning MiniMax-M3 leads
MiniMax-M3: 30.1 (#87), o4-mini: 24.6 (#162)
| Benchmark | MiniMax-M3 | o4-mini |
|---|---|---|
| SimpleBench | 45.8% | 38.7% |
| CritPt | 3.7% | 0.6% |
| Chess Puzzles | 14% | 26% |
| LMArena Hard Prompts | 1447 | 1351 |
| Mystery Game Puzzles | 8% | 5% |
| DTBench | 78.9% | 77.6% |
| LMCA | 33.7% | 26.5% |
| Epoch Capabilities Index | 146.95 | 145.64 |
| ForecastBench | 61.4 | 61.8 |
| ARC-AGI-2 | — | 6.1% |
| Kagi LLM Benchmark | — | 67.6% |
| NYT Connections (extended) | 65.1% | — |
| ARC-AGI-1 | — | 58.7% |
| EnigmaEval | — | 9.2% |
| Surface Evolver Bench | 55% | — |
Math Too close to call
MiniMax-M3: 40.0 (#95), o4-mini: 40.8 (#89)
| Benchmark | MiniMax-M3 | o4-mini |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 71.1% | 81.7% |
| LMArena Math | 1429 | 1389 |
| FrontierMath (Tiers 1-3) | — | 36.1% |
| FrontierMath Tier 4 | — | 4.9% |
| ProofBench | 18% | — |
| Omni-MATH | — | 72% |
| MATH Level 5 | — | 97.8% |
| FrontierMath (Feb 2025 set) | — | 24.8% |
| FrontierMath Tier 4 (v1) | — | 6.3% |
Knowledge MiniMax-M3 leads
MiniMax-M3: 58.4 (#35), o4-mini: 43.6 (#91)
| Benchmark | MiniMax-M3 | o4-mini |
|---|---|---|
| GPQA Diamond | 90.9% | 79.6% |
| LMArena Expert | 1461 | 1343 |
| Humanity's Last Exam | — | 18.1% |
| SimpleQA Verified | — | 19.6% |
| MMLU-Pro | — | 82% |
| Confabulations | — | 15.8% |
| Vectara Hallucination Rate | — | 18.6% |
| GPQA (HELM) | — | 73.5% |
Multimodal Too close to call
MiniMax-M3: 40.2 (#51), o4-mini: 40.2 (#49)
| Benchmark | MiniMax-M3 | o4-mini |
|---|---|---|
| LMArena Vision | 1253 | 1194 |
| GeoBench | — | 64% |
| VPCT | — | 57.5% |
| LMArena Document | 1435 | — |
Multilingual MiniMax-M3 leads
MiniMax-M3: 53.0 (#75), o4-mini: 47.0 (#154)
| Benchmark | MiniMax-M3 | o4-mini |
|---|---|---|
| LMArena Non-English | 1420 | 1337 |
| LMArena Chinese | 1463 | 1354 |
| LMArena French | 1447 | 1364 |
| LMArena German | 1426 | 1336 |
| LMArena Japanese | 1381 | 1308 |
| LMArena Korean | 1372 | 1312 |
| LMArena Russian | 1428 | 1334 |
| LMArena Spanish | 1432 | 1347 |
Instruction Following Too close to call
MiniMax-M3: 75.5 (#62), o4-mini: 75.2 (#68)
| Benchmark | MiniMax-M3 | o4-mini |
|---|---|---|
| LMArena Instruction Following | 1433 | 1321 |
| IFEval | — | 92.8% |
Long Context o4-mini leads
MiniMax-M3: 44.2 (#72), o4-mini: 45.5 (#33)
| Benchmark | MiniMax-M3 | o4-mini |
|---|---|---|
| LMArena Longer Query | 1445 | 1315 |
| Fiction.LiveBench | — | 77.8% |
Writing & Preference MiniMax-M3 leads
MiniMax-M3: 62.1 (#83), o4-mini: 54.0 (#152)
| Benchmark | MiniMax-M3 | o4-mini |
|---|---|---|
| LMArena Text | 1433 | 1353 |
| LMArena Creative Writing | 1404 | 1294 |
| LMArena Multi-Turn | 1442 | 1350 |
| Short-Story Creative Writing | — | 75% |
| WildBench | — | 85.4% |
| EQ-Bench 4 | 1150 | — |
Frequently asked questions
Is MiniMax-M3 better than o4-mini?
MiniMax-M3 is the stronger model overall, scoring 43.8 to 41.6 on the Noometry Index.
Which is cheaper, MiniMax-M3 or o4-mini?
MiniMax-M3 is cheaper. It lists at $0.30 per million input tokens and $1.20 per million output tokens; o4-mini lists at $1.10 and $4.40.
Is MiniMax-M3 or o4-mini better for coding?
They score almost the same on coding (41.8 vs 40.9); test both on your own repository before choosing.
Which has the bigger context window?
MiniMax-M3 does, with 1M tokens against 200K.
How many benchmarks do MiniMax-M3 and o4-mini share?
29 benchmarks have published results for both models. MiniMax-M3 has 41 scored results on Noometry and o4-mini has 60.