Model comparison
MiniMax-M2.5 vs Qwen3.5 397B-A17B
Qwen3.5 397B-A17B is the stronger model overall, scoring 46.0 to 38.3 on the Noometry Index. MiniMax-M2.5 costs 2.6× less per token, which makes it the better buy when Qwen3.5 397B-A17B's lead doesn't matter for your workload.
Last verified . 22 shared benchmarks.
MiniMax
Alibaba (Qwen)Summary
- They share 22 benchmarks with published results for both. MiniMax-M2.5 scores higher in 1 category and Qwen3.5 397B-A17B in 8 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3.5 397B-A17B leads 46.1 to 26.9.
- The biggest single-benchmark swing is NYT Connections (extended): 16.8% for MiniMax-M2.5 and 58.9% for Qwen3.5 397B-A17B.
- MiniMax-M2.5 is cheaper at $0.30 / $1.20 per million input/output tokens, against $0.60 / $3.60 for Qwen3.5 397B-A17B.
- Qwen3.5 397B-A17B accepts more context: 262K tokens versus 205K.
Side by side
| MiniMax-M2.5 | Qwen3.5 397B-A17B | |
|---|---|---|
| Provider | MiniMax | Alibaba (Qwen) |
| Noometry Index | 38.3 | 46.0 |
| Released | 2026-02-12 | 2026-02-01 |
| Weights | Open | Open |
| Context window | 205K | 262K |
| Max output | 131K | 66K |
| Input $ / M tokens | $0.30 | $0.60 |
| Output $ / M tokens | $1.20 | $3.60 |
| Results tracked | 33 | 36 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiniMax-M2.5 leads
MiniMax-M2.5: 48.1 (#58), Qwen3.5 397B-A17B: 42.0 (#114)
| Benchmark | MiniMax-M2.5 | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena WebDev | 1387 | 1400 |
| LMArena Coding | 1381 | 1465 |
| SWE-bench Verified (bash only) | 75.8% | — |
| SWE-bench Multilingual | 68.3% | — |
| ALE-Bench | 618.17 | — |
Agentic & Tool Use Qwen3.5 397B-A17B leads
MiniMax-M2.5: 30.4 (#77), Qwen3.5 397B-A17B: 33.3 (#53)
| Benchmark | MiniMax-M2.5 | Qwen3.5 397B-A17B |
|---|---|---|
| Terminal-Bench | 42.7% | — |
| APEX-Agents | — | 24.9% |
| τ²-bench Airline | — | 81.5% |
| τ²-bench Banking | — | 9.8% |
| τ²-bench Retail | — | 84.4% |
| τ²-bench Telecom | — | 97.8% |
| Vending-Bench 2 | -23.16 | — |
Reasoning Qwen3.5 397B-A17B leads
MiniMax-M2.5: 17.5 (#292), Qwen3.5 397B-A17B: 34.5 (#70)
| Benchmark | MiniMax-M2.5 | Qwen3.5 397B-A17B |
|---|---|---|
| Kagi LLM Benchmark | 55.2% | 73.7% |
| NYT Connections (extended) | 16.8% | 58.9% |
| LMArena Hard Prompts | 1372 | 1448 |
| Epoch Capabilities Index | 146.68 | 146.65 |
| ARC-AGI-2 | 4.9% | — |
| ARC-AGI-1 | 63.7% | — |
| Chess Puzzles | — | 13% |
| Thematic Generalization | — | 65.1% |
| Mystery Game Puzzles | — | 18% |
| DTBench | — | 87.5% |
| LMCA | — | 37.9% |
Math Qwen3.5 397B-A17B leads
MiniMax-M2.5: 26.9 (#253), Qwen3.5 397B-A17B: 46.1 (#73)
| Benchmark | MiniMax-M2.5 | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Math | 1378 | 1454 |
| FrontierMath (Tiers 1-3) | — | 31.2% |
| OTIS Mock AIME 2024-2025 | — | 88.9% |
| ProofBench | 4% | — |
Knowledge Qwen3.5 397B-A17B leads
MiniMax-M2.5: 39.2 (#135), Qwen3.5 397B-A17B: 53.3 (#58)
| Benchmark | MiniMax-M2.5 | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Expert | 1379 | 1462 |
| GPQA Diamond | — | 86.4% |
| Vectara Hallucination Rate | 9.1% | — |
Multimodal Not comparable
MiniMax-M2.5: —, Qwen3.5 397B-A17B: 40.7 (#44)
| Benchmark | MiniMax-M2.5 | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Vision | — | 1263 |
Multilingual Qwen3.5 397B-A17B leads
MiniMax-M2.5: 47.1 (#152), Qwen3.5 397B-A17B: 53.7 (#59)
| Benchmark | MiniMax-M2.5 | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Non-English | 1338 | 1430 |
| LMArena Chinese | 1393 | 1500 |
| LMArena French | 1362 | 1461 |
| LMArena German | 1362 | 1447 |
| LMArena Japanese | 1171 | 1426 |
| LMArena Korean | 1232 | 1384 |
| LMArena Russian | 1358 | 1429 |
| LMArena Spanish | 1354 | 1441 |
Instruction Following Qwen3.5 397B-A17B leads
MiniMax-M2.5: 71.5 (#148), Qwen3.5 397B-A17B: 75.0 (#77)
| Benchmark | MiniMax-M2.5 | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Instruction Following | 1353 | 1424 |
Long Context Qwen3.5 397B-A17B leads
MiniMax-M2.5: 37.5 (#216), Qwen3.5 397B-A17B: 44.1 (#74)
| Benchmark | MiniMax-M2.5 | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Longer Query | 1366 | 1442 |
| CL-bench | 11.4% | — |
| CL-bench Life | 6.3% | — |
Writing & Preference Qwen3.5 397B-A17B leads
MiniMax-M2.5: 53.9 (#153), Qwen3.5 397B-A17B: 62.3 (#79)
| Benchmark | MiniMax-M2.5 | Qwen3.5 397B-A17B |
|---|---|---|
| LMArena Text | 1359 | 1438 |
| LMArena Creative Writing | 1331 | 1401 |
| EQ-Bench Creative Writing | 1361 | 1478 |
| LMArena Multi-Turn | 1364 | 1446 |
Frequently asked questions
Is MiniMax-M2.5 better than Qwen3.5 397B-A17B?
Qwen3.5 397B-A17B is the stronger model overall, scoring 46.0 to 38.3 on the Noometry Index. MiniMax-M2.5 costs 2.6× less per token, which makes it the better buy when Qwen3.5 397B-A17B's lead doesn't matter for your workload.
Which is cheaper, MiniMax-M2.5 or Qwen3.5 397B-A17B?
MiniMax-M2.5 is cheaper. It lists at $0.30 per million input tokens and $1.20 per million output tokens; Qwen3.5 397B-A17B lists at $0.60 and $3.60.
Is MiniMax-M2.5 or Qwen3.5 397B-A17B better for coding?
MiniMax-M2.5 scores higher on coding benchmarks: 48.1 versus 42.0 in the Noometry coding category.
Which has the bigger context window?
Qwen3.5 397B-A17B does, with 262K tokens against 205K.
How many benchmarks do MiniMax-M2.5 and Qwen3.5 397B-A17B share?
22 benchmarks have published results for both models. MiniMax-M2.5 has 33 scored results on Noometry and Qwen3.5 397B-A17B has 36.