Model comparison
GPT-5 Mini vs MiniMax-M3
MiniMax-M3 is the stronger model overall, scoring 43.5 to 41.7 on the Noometry Index.
Last verified . 31 shared benchmarks.
MiniMaxSummary
- They share 31 benchmarks with published results for both. GPT-5 Mini scores higher in 3 categories and MiniMax-M3 in 7 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where MiniMax-M3 leads 58.4 to 45.5.
- The biggest single-benchmark swing is Chess Puzzles: 30% for GPT-5 Mini and 14% for MiniMax-M3.
- MiniMax-M3 is cheaper at $0.30 / $1.20 per million input/output tokens, against $0.25 / $2 for GPT-5 Mini.
- MiniMax-M3 accepts more context: 1M tokens versus 400K.
- MiniMax-M3 has downloadable open weights; the other is API-only.
Side by side
| GPT-5 Mini | MiniMax-M3 | |
|---|---|---|
| Provider | OpenAI | MiniMax |
| Noometry Index | 41.7 | 43.5 |
| Released | 2025-08-07 | 2026-06-01 |
| Weights | Proprietary | Open |
| Context window | 400K | 1M |
| Max output | 128K | 512K |
| Input $ / M tokens | $0.25 | $0.30 |
| Output $ / M tokens | $2 | $1.20 |
| Results tracked | 60 | 41 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiniMax-M3 leads
GPT-5 Mini: 40.0 (#134), MiniMax-M3: 41.5 (#111)
| Benchmark | GPT-5 Mini | MiniMax-M3 |
|---|---|---|
| SciCode | 39.2% | 47.1% |
| LMArena Coding | 1406 | 1470 |
| ALE-Bench | 799.77 | 640.02 |
| SWE-bench Verified | 64.7% | — |
| FrontierCode | — | 14.7% |
| SWE-bench Verified (bash only) | 59.8% | — |
| LMArena WebDev | — | 1482 |
| SWE-bench Multilingual | 39.7% | — |
| WeirdML | 52.7% | — |
| AlgoTune | 1.38 | — |
Agentic & Tool Use GPT-5 Mini leads
GPT-5 Mini: 31.0 (#64), MiniMax-M3: 22.5 (#121)
| Benchmark | GPT-5 Mini | MiniMax-M3 |
|---|---|---|
| Vending-Bench 2 | -31.18 | 2,158 |
| Terminal-Bench | 34.8% | — |
| APEX-Agents | — | 37.7% |
| Berkeley Function Calling Leaderboard | 55.5% | — |
| OSWorld 2.0 | — | 4.6% |
| GBAEval | — | 0.9% |
Reasoning MiniMax-M3 leads
GPT-5 Mini: 24.1 (#153), MiniMax-M3: 29.3 (#83)
| Benchmark | GPT-5 Mini | MiniMax-M3 |
|---|---|---|
| CritPt | 0% | 3.7% |
| Chess Puzzles | 30% | 14% |
| LMArena Hard Prompts | 1380 | 1447 |
| Mystery Game Puzzles | 10% | 8% |
| DTBench | 80.5% | 78.9% |
| LMCA | 34.2% | 33.7% |
| Epoch Capabilities Index | 145.51 | 147 |
| ForecastBench | 61 | 61.4 |
| ARC-AGI-2 | 4.4% | — |
| SimpleBench | — | 45.8% |
| Kagi LLM Benchmark | 70.3% | — |
| NYT Connections (extended) | — | 65.1% |
| ARC-AGI-1 | 54.3% | — |
| EnigmaEval | 8.2% | — |
| Surface Evolver Bench | — | 55% |
Math GPT-5 Mini leads
GPT-5 Mini: 46.3 (#59), MiniMax-M3: 39.1 (#83)
| Benchmark | GPT-5 Mini | MiniMax-M3 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 86.7% | 71.1% |
| ProofBench | 9% | 18% |
| LMArena Math | 1377 | 1432 |
| FrontierMath (Tiers 1-3) | 46.7% | — |
| FrontierMath Tier 4 | 12.2% | — |
| Omni-MATH | 72.2% | — |
| MATH Level 5 | 97.8% | — |
| FrontierMath (Feb 2025 set) | 27.2% | — |
| FrontierMath Tier 4 (v1) | 6.3% | — |
Knowledge MiniMax-M3 leads
GPT-5 Mini: 45.5 (#80), MiniMax-M3: 58.4 (#32)
| Benchmark | GPT-5 Mini | MiniMax-M3 |
|---|---|---|
| GPQA Diamond | 75% | 90.9% |
| LMArena Expert | 1377 | 1462 |
| Humanity's Last Exam | 19.4% | — |
| SimpleQA Verified | 21.6% | — |
| MMLU-Pro | 83.5% | — |
| Confabulations | 13.3% | — |
| Vectara Hallucination Rate | 12.9% | — |
| GPQA (HELM) | 75.6% | — |
Multimodal MiniMax-M3 leads
GPT-5 Mini: 35.0 (#80), MiniMax-M3: 39.5 (#45)
| Benchmark | GPT-5 Mini | MiniMax-M3 |
|---|---|---|
| LMArena Vision | 1202 | 1254 |
| VPCT | 40.2% | — |
| LMArena Document | — | 1435 |
Multilingual MiniMax-M3 leads
GPT-5 Mini: 49.1 (#131), MiniMax-M3: 53.2 (#70)
| Benchmark | GPT-5 Mini | MiniMax-M3 |
|---|---|---|
| LMArena Non-English | 1363 | 1420 |
| LMArena Chinese | 1387 | 1465 |
| LMArena French | 1385 | 1445 |
| LMArena German | 1366 | 1424 |
| LMArena Japanese | 1342 | 1381 |
| LMArena Korean | 1306 | 1372 |
| LMArena Russian | 1362 | 1430 |
| LMArena Spanish | 1355 | 1431 |
Instruction Following Too close to call
GPT-5 Mini: 76.4 (#43), MiniMax-M3: 75.7 (#56)
| Benchmark | GPT-5 Mini | MiniMax-M3 |
|---|---|---|
| LMArena Instruction Following | 1357 | 1433 |
| IFEval | 92.7% | — |
Long Context MiniMax-M3 leads
GPT-5 Mini: 41.9 (#127), MiniMax-M3: 44.2 (#68)
| Benchmark | GPT-5 Mini | MiniMax-M3 |
|---|---|---|
| LMArena Longer Query | 1355 | 1445 |
| Fiction.LiveBench | 69.4% | — |
Writing & Preference MiniMax-M3 leads
GPT-5 Mini: 55.4 (#143), MiniMax-M3: 62.3 (#79)
| Benchmark | GPT-5 Mini | MiniMax-M3 |
|---|---|---|
| LMArena Text | 1373 | 1433 |
| LMArena Creative Writing | 1325 | 1405 |
| LMArena Multi-Turn | 1363 | 1441 |
| Short-Story Creative Writing | 83.1% | — |
| EQ-Bench Creative Writing | 1313 | — |
| WildBench | 85.5% | — |
| EQ-Bench 4 | — | 1150 |
Frequently asked questions
Is GPT-5 Mini better than MiniMax-M3?
MiniMax-M3 is the stronger model overall, scoring 43.5 to 41.7 on the Noometry Index.
Which is cheaper, GPT-5 Mini or MiniMax-M3?
MiniMax-M3 is cheaper. It lists at $0.30 per million input tokens and $1.20 per million output tokens; GPT-5 Mini lists at $0.25 and $2.
Is GPT-5 Mini or MiniMax-M3 better for coding?
MiniMax-M3 scores higher on coding benchmarks: 41.5 versus 40.0 in the Noometry coding category.
Which has the bigger context window?
MiniMax-M3 does, with 1M tokens against 400K.
How many benchmarks do GPT-5 Mini and MiniMax-M3 share?
31 benchmarks have published results for both models. GPT-5 Mini has 60 scored results on Noometry and MiniMax-M3 has 41.