Model comparison
GPT-4.1 vs MiniMax-M2.5
MiniMax-M2.5 is the stronger model overall, scoring 38.1 to 35.7 on the Noometry Index.
Last verified . 25 shared benchmarks.
MiniMaxSummary
- They share 25 benchmarks with published results for both. GPT-4.1 scores higher in 4 categories and MiniMax-M2.5 in 5 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in coding, where MiniMax-M2.5 leads 47.9 to 34.2.
- The biggest single-benchmark swing is ARC-AGI-1: 5.5% for GPT-4.1 and 63.7% for MiniMax-M2.5.
- MiniMax-M2.5 is cheaper at $0.30 / $1.20 per million input/output tokens, against $2 / $8 for GPT-4.1.
- GPT-4.1 accepts more context: 1.05M tokens versus 205K.
- MiniMax-M2.5 has downloadable open weights; the other is API-only.
Side by side
| GPT-4.1 | MiniMax-M2.5 | |
|---|---|---|
| Provider | OpenAI | MiniMax |
| Noometry Index | 35.7 | 38.1 |
| Released | 2025-04-14 | 2026-02-12 |
| Weights | Proprietary | Open |
| Context window | 1.05M | 205K |
| Max output | 33K | 131K |
| Input $ / M tokens | $2 | $0.30 |
| Output $ / M tokens | $8 | $1.20 |
| Results tracked | 52 | 33 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiniMax-M2.5 leads
GPT-4.1: 34.2 (#228), MiniMax-M2.5: 47.9 (#53)
| Benchmark | GPT-4.1 | MiniMax-M2.5 |
|---|---|---|
| SWE-bench Verified (bash only) | 39.6% | 75.8% |
| LMArena Coding | 1391 | 1380 |
| ALE-Bench | 558.1 | 618.17 |
| SWE-bench Verified | 48.5% | — |
| Aider Polyglot | 52.4% | — |
| LMArena WebDev | — | 1386 |
| SWE-bench Multilingual | — | 68.3% |
| WeirdML | 39% | — |
| CadEval | 42% | — |
Agentic & Tool Use GPT-4.1 leads
GPT-4.1: 34.4 (#40), MiniMax-M2.5: 30.2 (#70)
| Benchmark | GPT-4.1 | MiniMax-M2.5 |
|---|---|---|
| Terminal-Bench | — | 42.7% |
| Berkeley Function Calling Leaderboard | 54% | — |
| Vending-Bench 2 | — | -23.16 |
Reasoning MiniMax-M2.5 leads
GPT-4.1: 11.7 (#334), MiniMax-M2.5: 17.6 (#281)
| Benchmark | GPT-4.1 | MiniMax-M2.5 |
|---|---|---|
| ARC-AGI-2 | 0.4% | 4.9% |
| Kagi LLM Benchmark | 52.3% | 55.2% |
| ARC-AGI-1 | 5.5% | 63.7% |
| LMArena Hard Prompts | 1384 | 1372 |
| Epoch Capabilities Index | 136.79 | 146.68 |
| SimpleBench | 27% | — |
| NYT Connections (extended) | — | 16.8% |
| Chess Puzzles | 6% | — |
| EnigmaEval | 2.2% | — |
| DTBench | 68.3% | — |
| LMCA | 25.6% | — |
| ForecastBench | 61.5 | — |
Math MiniMax-M2.5 leads
GPT-4.1: 21.4 (#269), MiniMax-M2.5: 25.8 (#242)
| Benchmark | GPT-4.1 | MiniMax-M2.5 |
|---|---|---|
| LMArena Math | 1369 | 1378 |
| FrontierMath (Tiers 1-3) | 6% | — |
| OTIS Mock AIME 2024-2025 | 38.3% | — |
| ProofBench | — | 4% |
| Omni-MATH | 47.1% | — |
| MATH Level 5 | 83% | — |
| FrontierMath (Feb 2025 set) | 5.5% | — |
| FrontierMath Tier 4 (v1) | 0% | — |
Knowledge MiniMax-M2.5 leads
GPT-4.1: 37.0 (#152), MiniMax-M2.5: 38.6 (#132)
| Benchmark | GPT-4.1 | MiniMax-M2.5 |
|---|---|---|
| Vectara Hallucination Rate | 5.6% | 9.1% |
| LMArena Expert | 1364 | 1378 |
| GPQA Diamond | 66.9% | — |
| Humanity's Last Exam | 5.4% | — |
| SimpleQA Verified | 31.1% | — |
| MMLU-Pro | 81.1% | — |
| GPQA (HELM) | 65.9% | — |
Multimodal Not comparable
GPT-4.1: 37.5 (#66), MiniMax-M2.5: —
| Benchmark | GPT-4.1 | MiniMax-M2.5 |
|---|---|---|
| LMArena Vision | 1210 | — |
| GeoBench | 72% | — |
Multilingual GPT-4.1 leads
GPT-4.1: 49.6 (#128), MiniMax-M2.5: 47.3 (#147)
| Benchmark | GPT-4.1 | MiniMax-M2.5 |
|---|---|---|
| LMArena Non-English | 1370 | 1338 |
| LMArena Chinese | 1382 | 1392 |
| LMArena French | 1383 | 1363 |
| LMArena German | 1380 | 1361 |
| LMArena Japanese | 1320 | 1169 |
| LMArena Korean | 1339 | 1232 |
| LMArena Russian | 1377 | 1359 |
| LMArena Spanish | 1375 | 1354 |
Instruction Following Too close to call
GPT-4.1: 71.5 (#148), MiniMax-M2.5: 71.7 (#143)
| Benchmark | GPT-4.1 | MiniMax-M2.5 |
|---|---|---|
| LMArena Instruction Following | 1367 | 1353 |
| IFEval | 83.8% | — |
Long Context GPT-4.1 leads
GPT-4.1: 40.0 (#158), MiniMax-M2.5: 37.4 (#211)
| Benchmark | GPT-4.1 | MiniMax-M2.5 |
|---|---|---|
| LMArena Longer Query | 1385 | 1367 |
| Fiction.LiveBench | 63.9% | — |
| CL-bench | — | 11.4% |
| CL-bench Life | — | 6.3% |
Writing & Preference GPT-4.1 leads
GPT-4.1: 57.8 (#120), MiniMax-M2.5: 54.1 (#149)
| Benchmark | GPT-4.1 | MiniMax-M2.5 |
|---|---|---|
| LMArena Text | 1383 | 1359 |
| LMArena Creative Writing | 1363 | 1331 |
| EQ-Bench Creative Writing | 1420 | 1361 |
| LMArena Multi-Turn | 1398 | 1364 |
| WildBench | 85.4% | — |
Frequently asked questions
Is GPT-4.1 better than MiniMax-M2.5?
MiniMax-M2.5 is the stronger model overall, scoring 38.1 to 35.7 on the Noometry Index.
Which is cheaper, GPT-4.1 or MiniMax-M2.5?
MiniMax-M2.5 is cheaper. It lists at $0.30 per million input tokens and $1.20 per million output tokens; GPT-4.1 lists at $2 and $8.
Is GPT-4.1 or MiniMax-M2.5 better for coding?
MiniMax-M2.5 scores higher on coding benchmarks: 47.9 versus 34.2 in the Noometry coding category.
Which has the bigger context window?
GPT-4.1 does, with 1.05M tokens against 205K.
How many benchmarks do GPT-4.1 and MiniMax-M2.5 share?
25 benchmarks have published results for both models. GPT-4.1 has 52 scored results on Noometry and MiniMax-M2.5 has 33.