Model comparison
GPT-5 vs MiniMax-M3
GPT-5 is the stronger model overall, scoring 50.9 to 43.8 on the Noometry Index. MiniMax-M3 costs 6.5× less per token, which makes it the better buy when GPT-5's lead doesn't matter for your workload.
Last verified . 32 shared benchmarks.
MiniMaxSummary
- They share 32 benchmarks with published results for both. GPT-5 scores higher in 7 categories and MiniMax-M3 in 3 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in long context, where GPT-5 leads 69.5 to 44.2.
- The biggest single-benchmark swing is Chess Puzzles: 37% for GPT-5 and 14% for MiniMax-M3.
- MiniMax-M3 is cheaper at $0.30 / $1.20 per million input/output tokens, against $1.25 / $10 for GPT-5.
- 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 | MiniMax-M3 | |
|---|---|---|
| Provider | OpenAI | MiniMax |
| Noometry Index | 50.9 | 43.8 |
| Released | 2025-08-07 | 2026-06-01 |
| Weights | Proprietary | Open |
| Context window | 400K | 1M |
| Max output | 128K | 512K |
| Input $ / M tokens | $1.25 | $0.30 |
| Output $ / M tokens | $10 | $1.20 |
| Results tracked | 69 | 41 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5 leads
GPT-5: 50.3 (#47), MiniMax-M3: 41.8 (#118)
| Benchmark | GPT-5 | MiniMax-M3 |
|---|---|---|
| LMArena WebDev | 1418 | 1482 |
| SciCode | 42.9% | 47.1% |
| LMArena Coding | 1436 | 1469 |
| ALE-Bench | 1,162 | 640.02 |
| SWE-bench Verified | 73.6% | — |
| FrontierCode | — | 14.7% |
| SWE-bench Verified (bash only) | 65% | — |
| Aider Polyglot | 88% | — |
| GSO | 6.9% | — |
| WeirdML | 60.7% | — |
| AlgoTune | 1.67 | — |
Agentic & Tool Use GPT-5 leads
GPT-5: 33.1 (#56), MiniMax-M3: 22.6 (#130)
| Benchmark | GPT-5 | MiniMax-M3 |
|---|---|---|
| Terminal-Bench | 49.6% | — |
| APEX-Agents | — | 37.7% |
| OSWorld 2.0 | — | 4.6% |
| GDPval | 34.8% | — |
| Remote Labor Index | 1.7% | — |
| DeepResearch Bench | 49.6% | — |
| BALROG | 32.8% | — |
| GBAEval | — | 0.9% |
| LMArena Search | 1133 | — |
| METR Time Horizons | 69.6% | — |
| Vending-Bench 2 | — | 2,158 |
Reasoning GPT-5 leads
GPT-5: 38.3 (#64), MiniMax-M3: 30.1 (#87)
| Benchmark | GPT-5 | MiniMax-M3 |
|---|---|---|
| SimpleBench | 56.7% | 45.8% |
| CritPt | 12.6% | 3.7% |
| Chess Puzzles | 37% | 14% |
| LMArena Hard Prompts | 1416 | 1447 |
| Mystery Game Puzzles | 23% | 8% |
| DTBench | 90.7% | 78.9% |
| LMCA | 40% | 33.7% |
| Epoch Capabilities Index | 150 | 146.95 |
| ForecastBench | 61.4 | 61.4 |
| ARC-AGI-2 | 9.9% | — |
| Kagi LLM Benchmark | 72.7% | — |
| NYT Connections (extended) | — | 65.1% |
| ARC-AGI-1 | 65.7% | — |
| EnigmaEval | 10.5% | — |
| EBR-Bench | 12.7% | — |
| Surface Evolver Bench | — | 55% |
Math GPT-5 leads
GPT-5: 55.0 (#44), MiniMax-M3: 40.0 (#95)
| Benchmark | GPT-5 | MiniMax-M3 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 91.4% | 71.1% |
| ProofBench | 18% | 18% |
| LMArena Math | 1407 | 1429 |
| FrontierMath (Tiers 1-3) | 55.4% | — |
| FrontierMath Tier 4 | 22% | — |
| Omni-MATH | 64.7% | — |
| MATH Level 5 | 98.1% | — |
| FrontierMath (Feb 2025 set) | 32.4% | — |
| FrontierMath Tier 4 (v1) | 12.5% | — |
Knowledge MiniMax-M3 leads
GPT-5: 56.6 (#43), MiniMax-M3: 58.4 (#35)
| Benchmark | GPT-5 | MiniMax-M3 |
|---|---|---|
| GPQA Diamond | 86.2% | 90.9% |
| LMArena Expert | 1419 | 1461 |
| Humanity's Last Exam | 25.3% | — |
| SimpleQA Verified | 50.1% | — |
| MMLU-Pro | 86.3% | — |
| Confabulations | 10.3% | — |
| Vectara Hallucination Rate | 14.7% | — |
| GPQA (HELM) | 79.2% | — |
Multimodal GPT-5 leads
GPT-5: 46.8 (#13), MiniMax-M3: 40.2 (#51)
| Benchmark | GPT-5 | MiniMax-M3 |
|---|---|---|
| LMArena Vision | 1232 | 1253 |
| GeoBench | 81% | — |
| VPCT | 66% | — |
| LMArena Document | — | 1435 |
Multilingual MiniMax-M3 leads
GPT-5: 51.4 (#110), MiniMax-M3: 53.0 (#75)
| Benchmark | GPT-5 | MiniMax-M3 |
|---|---|---|
| LMArena Non-English | 1397 | 1420 |
| LMArena Chinese | 1422 | 1463 |
| LMArena French | 1410 | 1447 |
| LMArena German | 1416 | 1426 |
| LMArena Japanese | 1409 | 1381 |
| LMArena Korean | 1360 | 1372 |
| LMArena Russian | 1406 | 1428 |
| LMArena Spanish | 1399 | 1432 |
Instruction Following MiniMax-M3 leads
GPT-5: 73.8 (#113), MiniMax-M3: 75.5 (#62)
| Benchmark | GPT-5 | MiniMax-M3 |
|---|---|---|
| LMArena Instruction Following | 1388 | 1433 |
| IFEval | 87.5% | — |
Long Context GPT-5 leads
GPT-5: 69.5 (#2), MiniMax-M3: 44.2 (#72)
| Benchmark | GPT-5 | MiniMax-M3 |
|---|---|---|
| LMArena Longer Query | 1399 | 1445 |
| Fiction.LiveBench | 97.2% | — |
Writing & Preference GPT-5 leads
GPT-5: 63.4 (#65), MiniMax-M3: 62.1 (#83)
| Benchmark | GPT-5 | MiniMax-M3 |
|---|---|---|
| LMArena Text | 1406 | 1433 |
| LMArena Creative Writing | 1365 | 1404 |
| LMArena Multi-Turn | 1426 | 1442 |
| Short-Story Creative Writing | 86% | — |
| EQ-Bench Creative Writing | 1627 | — |
| WildBench | 85.7% | — |
| EQ-Bench 4 | — | 1150 |
Frequently asked questions
Is GPT-5 better than MiniMax-M3?
GPT-5 is the stronger model overall, scoring 50.9 to 43.8 on the Noometry Index. MiniMax-M3 costs 6.5× less per token, which makes it the better buy when GPT-5's lead doesn't matter for your workload.
Which is cheaper, GPT-5 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 lists at $1.25 and $10.
Is GPT-5 or MiniMax-M3 better for coding?
GPT-5 scores higher on coding benchmarks: 50.3 versus 41.8 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 and MiniMax-M3 share?
32 benchmarks have published results for both models. GPT-5 has 69 scored results on Noometry and MiniMax-M3 has 41.