Model comparison
MiniMax-M2 vs Qwen3.5 27B
Qwen3.5 27B is the stronger model overall, scoring 41.4 to 36.9 on the Noometry Index. MiniMax-M2 costs 1.6× less per token, which makes it the better buy when Qwen3.5 27B's lead doesn't matter for your workload.
Last verified . 18 shared benchmarks.
MiniMax
Alibaba (Qwen)Summary
- They share 18 benchmarks with published results for both. MiniMax-M2 scores higher in 1 category and Qwen3.5 27B in 7 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Qwen3.5 27B leads 26.6 to 18.6.
- The biggest single-benchmark swing is NYT Connections (extended): 14.8% for MiniMax-M2 and 47.9% for Qwen3.5 27B.
- MiniMax-M2 is cheaper at $0.30 / $1.20 per million input/output tokens, against $0.30 / $2.40 for Qwen3.5 27B.
- Qwen3.5 27B accepts more context: 262K tokens versus 205K.
Side by side
| MiniMax-M2 | Qwen3.5 27B | |
|---|---|---|
| Provider | MiniMax | Alibaba (Qwen) |
| Noometry Index | 36.9 | 41.4 |
| Released | 2025-10-27 | 2026-02-23 |
| Weights | Open | Open |
| Context window | 205K | 262K |
| Max output | 131K | 66K |
| Input $ / M tokens | $0.30 | $0.30 |
| Output $ / M tokens | $1.20 | $2.40 |
| Results tracked | 21 | 28 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
MiniMax-M2: 39.1 (#151), Qwen3.5 27B: 38.6 (#159)
| Benchmark | MiniMax-M2 | Qwen3.5 27B |
|---|---|---|
| LMArena WebDev | 1297 | 1357 |
| LMArena Coding | 1370 | 1427 |
| SWE-bench Verified (bash only) | 61% | — |
| WeirdML | — | 39.5% |
| ALE-Bench | — | 349.45 |
Agentic & Tool Use Not comparable
MiniMax-M2: 24.9 (#102), Qwen3.5 27B: —
| Benchmark | MiniMax-M2 | Qwen3.5 27B |
|---|---|---|
| Vending-Bench 2 | 160.6 | 201.98 |
| Terminal-Bench | 30% | — |
Reasoning Qwen3.5 27B leads
MiniMax-M2: 18.6 (#256), Qwen3.5 27B: 26.6 (#116)
| Benchmark | MiniMax-M2 | Qwen3.5 27B |
|---|---|---|
| NYT Connections (extended) | 14.8% | 47.9% |
| LMArena Hard Prompts | 1357 | 1414 |
| Kagi LLM Benchmark | 57.8% | — |
| Thematic Generalization | — | 45.5% |
| DTBench | — | 82.4% |
| LMCA | — | 34% |
Math Qwen3.5 27B leads
MiniMax-M2: 34.6 (#158), Qwen3.5 27B: 36.8 (#103)
| Benchmark | MiniMax-M2 | Qwen3.5 27B |
|---|---|---|
| LMArena Math | 1350 | 1428 |
| MathArena Final-Answer Competitions | — | 56.7% |
Knowledge Qwen3.5 27B leads
MiniMax-M2: 36.3 (#160), Qwen3.5 27B: 37.4 (#149)
| Benchmark | MiniMax-M2 | Qwen3.5 27B |
|---|---|---|
| LMArena Expert | 1336 | 1428 |
| Vectara Hallucination Rate | — | 12.1% |
Multimodal Not comparable
MiniMax-M2: —, Qwen3.5 27B: 38.7 (#56)
| Benchmark | MiniMax-M2 | Qwen3.5 27B |
|---|---|---|
| LMArena Vision | — | 1240 |
Multilingual Qwen3.5 27B leads
MiniMax-M2: 45.6 (#165), Qwen3.5 27B: 51.0 (#112)
| Benchmark | MiniMax-M2 | Qwen3.5 27B |
|---|---|---|
| LMArena Non-English | 1314 | 1390 |
| LMArena Chinese | 1367 | 1478 |
| LMArena French | 1336 | 1409 |
| LMArena German | 1366 | 1393 |
| LMArena Russian | 1332 | 1389 |
| LMArena Spanish | 1326 | 1408 |
| LMArena Japanese | — | 1345 |
| LMArena Korean | — | 1359 |
Instruction Following Qwen3.5 27B leads
MiniMax-M2: 70.4 (#160), Qwen3.5 27B: 73.7 (#114)
| Benchmark | MiniMax-M2 | Qwen3.5 27B |
|---|---|---|
| LMArena Instruction Following | 1329 | 1393 |
Long Context Qwen3.5 27B leads
MiniMax-M2: 40.5 (#148), Qwen3.5 27B: 43.1 (#103)
| Benchmark | MiniMax-M2 | Qwen3.5 27B |
|---|---|---|
| LMArena Longer Query | 1332 | 1413 |
Writing & Preference Qwen3.5 27B leads
MiniMax-M2: 53.3 (#157), Qwen3.5 27B: 59.5 (#106)
| Benchmark | MiniMax-M2 | Qwen3.5 27B |
|---|---|---|
| LMArena Text | 1340 | 1409 |
| LMArena Creative Writing | 1288 | 1362 |
| LMArena Multi-Turn | 1362 | 1410 |
Frequently asked questions
Is MiniMax-M2 better than Qwen3.5 27B?
Qwen3.5 27B is the stronger model overall, scoring 41.4 to 36.9 on the Noometry Index. MiniMax-M2 costs 1.6× less per token, which makes it the better buy when Qwen3.5 27B's lead doesn't matter for your workload.
Which is cheaper, MiniMax-M2 or Qwen3.5 27B?
MiniMax-M2 is cheaper. It lists at $0.30 per million input tokens and $1.20 per million output tokens; Qwen3.5 27B lists at $0.30 and $2.40.
Is MiniMax-M2 or Qwen3.5 27B better for coding?
They score almost the same on coding (39.1 vs 38.6); test both on your own repository before choosing.
Which has the bigger context window?
Qwen3.5 27B does, with 262K tokens against 205K.
How many benchmarks do MiniMax-M2 and Qwen3.5 27B share?
18 benchmarks have published results for both models. MiniMax-M2 has 21 scored results on Noometry and Qwen3.5 27B has 28.