Model comparison
MiniMax-M2.5 vs Qwen3.7 Max
Qwen3.7 Max is the stronger model overall, scoring 51.4 to 38.1 on the Noometry Index. MiniMax-M2.5 costs 7.1× less per token, which makes it the better buy when Qwen3.7 Max's lead doesn't matter for your workload.
Last verified . 17 shared benchmarks.
MiniMax
Alibaba (Qwen)Summary
- They share 17 benchmarks with published results for both. MiniMax-M2.5 scores higher in 1 category and Qwen3.7 Max in 8 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3.7 Max leads 62.3 to 25.8.
- The biggest single-benchmark swing is NYT Connections (extended): 16.8% for MiniMax-M2.5 and 85.1% for Qwen3.7 Max.
- MiniMax-M2.5 is cheaper at $0.30 / $1.20 per million input/output tokens, against $2.50 / $7.50 for Qwen3.7 Max.
- Qwen3.7 Max accepts more context: 1M tokens versus 205K.
- MiniMax-M2.5 has downloadable open weights; the other is API-only.
Side by side
| MiniMax-M2.5 | Qwen3.7 Max | |
|---|---|---|
| Provider | MiniMax | Alibaba (Qwen) |
| Noometry Index | 38.1 | 51.4 |
| Released | 2026-02-12 | 2026-05-19 |
| Weights | Open | Proprietary |
| Context window | 205K | 1M |
| Max output | 131K | 131K |
| Input $ / M tokens | $0.30 | $2.50 |
| Output $ / M tokens | $1.20 | $7.50 |
| Results tracked | 33 | 33 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.7 Max leads
MiniMax-M2.5: 47.9 (#53), Qwen3.7 Max: 50.2 (#43)
| Benchmark | MiniMax-M2.5 | Qwen3.7 Max |
|---|---|---|
| LMArena WebDev | 1386 | 1515 |
| LMArena Coding | 1380 | 1498 |
| ALE-Bench | 618.17 | 1,189 |
| SWE-bench Verified | — | 77.3% |
| SWE-bench Verified (bash only) | 75.8% | — |
| SWE-bench Multilingual | 68.3% | — |
| SciCode | — | 48.8% |
Agentic & Tool Use MiniMax-M2.5 leads
MiniMax-M2.5: 30.2 (#70), Qwen3.7 Max: 21.8 (#129)
| Benchmark | MiniMax-M2.5 | Qwen3.7 Max |
|---|---|---|
| Terminal-Bench | 42.7% | — |
| GBAEval | — | 0.4% |
| Vending-Bench 2 | -23.16 | — |
Reasoning Qwen3.7 Max leads
MiniMax-M2.5: 17.6 (#281), Qwen3.7 Max: 48.4 (#36)
| Benchmark | MiniMax-M2.5 | Qwen3.7 Max |
|---|---|---|
| NYT Connections (extended) | 16.8% | 85.1% |
| LMArena Hard Prompts | 1372 | 1483 |
| Epoch Capabilities Index | 146.68 | 153.74 |
| ARC-AGI-2 | 4.9% | — |
| SimpleBench | — | 70.4% |
| Kagi LLM Benchmark | 55.2% | — |
| ARC-AGI-1 | 63.7% | — |
| CritPt | — | 13.4% |
| Chess Puzzles | — | 19% |
| EBR-Bench | — | 9.5% |
| Mystery Game Puzzles | — | 32% |
| DTBench | — | 92.3% |
| LMCA | — | 44% |
Math Qwen3.7 Max leads
MiniMax-M2.5: 25.8 (#242), Qwen3.7 Max: 62.3 (#26)
| Benchmark | MiniMax-M2.5 | Qwen3.7 Max |
|---|---|---|
| ProofBench | 4% | 26% |
| LMArena Math | 1378 | 1488 |
| FrontierMath (Tiers 1-3) | — | 64.6% |
| FrontierMath Tier 4 | — | 34.1% |
| OTIS Mock AIME 2024-2025 | — | 95.6% |
Knowledge Qwen3.7 Max leads
MiniMax-M2.5: 38.6 (#132), Qwen3.7 Max: 61.7 (#24)
| Benchmark | MiniMax-M2.5 | Qwen3.7 Max |
|---|---|---|
| LMArena Expert | 1378 | 1488 |
| GPQA Diamond | — | 90.9% |
| SimpleQA Verified | — | 55.8% |
| Vectara Hallucination Rate | 9.1% | — |
Multilingual Qwen3.7 Max leads
MiniMax-M2.5: 47.3 (#147), Qwen3.7 Max: 57.1 (#13)
| Benchmark | MiniMax-M2.5 | Qwen3.7 Max |
|---|---|---|
| LMArena Non-English | 1338 | 1475 |
| LMArena Chinese | 1392 | 1533 |
| LMArena Russian | 1359 | 1482 |
| LMArena French | 1363 | — |
| LMArena German | 1361 | — |
| LMArena Japanese | 1169 | — |
| LMArena Korean | 1232 | — |
| LMArena Spanish | 1354 | — |
Instruction Following Qwen3.7 Max leads
MiniMax-M2.5: 71.7 (#143), Qwen3.7 Max: 76.9 (#34)
| Benchmark | MiniMax-M2.5 | Qwen3.7 Max |
|---|---|---|
| LMArena Instruction Following | 1353 | 1460 |
Long Context Qwen3.7 Max leads
MiniMax-M2.5: 37.4 (#211), Qwen3.7 Max: 45.4 (#34)
| Benchmark | MiniMax-M2.5 | Qwen3.7 Max |
|---|---|---|
| LMArena Longer Query | 1367 | 1484 |
| CL-bench | 11.4% | — |
| CL-bench Life | 6.3% | — |
Writing & Preference Qwen3.7 Max leads
MiniMax-M2.5: 54.1 (#149), Qwen3.7 Max: 65.2 (#52)
| Benchmark | MiniMax-M2.5 | Qwen3.7 Max |
|---|---|---|
| LMArena Text | 1359 | 1476 |
| LMArena Creative Writing | 1331 | 1450 |
| LMArena Multi-Turn | 1364 | 1482 |
| EQ-Bench Creative Writing | 1361 | — |
| EQ-Bench 4 | — | 1110 |
Frequently asked questions
Is MiniMax-M2.5 better than Qwen3.7 Max?
Qwen3.7 Max is the stronger model overall, scoring 51.4 to 38.1 on the Noometry Index. MiniMax-M2.5 costs 7.1× less per token, which makes it the better buy when Qwen3.7 Max's lead doesn't matter for your workload.
Which is cheaper, MiniMax-M2.5 or Qwen3.7 Max?
MiniMax-M2.5 is cheaper. It lists at $0.30 per million input tokens and $1.20 per million output tokens; Qwen3.7 Max lists at $2.50 and $7.50.
Is MiniMax-M2.5 or Qwen3.7 Max better for coding?
Qwen3.7 Max scores higher on coding benchmarks: 50.2 versus 47.9 in the Noometry coding category.
Which has the bigger context window?
Qwen3.7 Max does, with 1M tokens against 205K.
How many benchmarks do MiniMax-M2.5 and Qwen3.7 Max share?
17 benchmarks have published results for both models. MiniMax-M2.5 has 33 scored results on Noometry and Qwen3.7 Max has 33.