Model comparison
MiniMax-M2.5 vs Qwen3 235B-A22B
Qwen3 235B-A22B is the stronger model overall, scoring 43.5 to 38.3 on the Noometry Index. MiniMax-M2.5 costs 2.3× less per token, which makes it the better buy when Qwen3 235B-A22B's lead doesn't matter for your workload.
Last verified . 24 shared benchmarks.
MiniMax
Alibaba (Qwen)Summary
- They share 24 benchmarks with published results for both. MiniMax-M2.5 scores higher in 2 categories and Qwen3 235B-A22B in 7 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3 235B-A22B leads 50.4 to 26.9.
- The biggest single-benchmark swing is ARC-AGI-1: 63.7% for MiniMax-M2.5 and 11% for Qwen3 235B-A22B.
- MiniMax-M2.5 is cheaper at $0.30 / $1.20 per million input/output tokens, against $0.70 / $2.80 for Qwen3 235B-A22B.
- MiniMax-M2.5 accepts more context: 205K tokens versus 131K.
Side by side
| MiniMax-M2.5 | Qwen3 235B-A22B | |
|---|---|---|
| Provider | MiniMax | Alibaba (Qwen) |
| Noometry Index | 38.3 | 43.5 |
| Released | 2026-02-12 | 2025-04 |
| Weights | Open | Open |
| Context window | 205K | 131K |
| Max output | 131K | 16K |
| Input $ / M tokens | $0.30 | $0.70 |
| Output $ / M tokens | $1.20 | $2.80 |
| Results tracked | 33 | 49 |
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 235B-A22B: 44.3 (#75)
| Benchmark | MiniMax-M2.5 | Qwen3 235B-A22B |
|---|---|---|
| LMArena Coding | 1381 | 1445 |
| SWE-bench Verified (bash only) | 75.8% | — |
| Aider Polyglot | — | 59.6% |
| LMArena WebDev | 1387 | — |
| SWE-bench Multilingual | 68.3% | — |
| SciCode | — | 42.4% |
| WeirdML | — | 41% |
| ALE-Bench | 618.17 | — |
Agentic & Tool Use Qwen3 235B-A22B leads
MiniMax-M2.5: 30.4 (#77), Qwen3 235B-A22B: 33.9 (#51)
| Benchmark | MiniMax-M2.5 | Qwen3 235B-A22B |
|---|---|---|
| Vending-Bench 2 | -23.16 | -11.34 |
| Terminal-Bench | 42.7% | — |
| Berkeley Function Calling Leaderboard | — | 52.1% |
Reasoning MiniMax-M2.5 leads
MiniMax-M2.5: 17.5 (#292), Qwen3 235B-A22B: 15.7 (#311)
| Benchmark | MiniMax-M2.5 | Qwen3 235B-A22B |
|---|---|---|
| ARC-AGI-2 | 4.9% | 1.3% |
| Kagi LLM Benchmark | 55.2% | 69.4% |
| ARC-AGI-1 | 63.7% | 11% |
| LMArena Hard Prompts | 1372 | 1433 |
| Epoch Capabilities Index | 146.68 | 143.85 |
| SimpleBench | — | 31% |
| NYT Connections (extended) | 16.8% | — |
| CritPt | — | 0% |
| Chess Puzzles | — | 12% |
| Mystery Game Puzzles | — | 9% |
| DTBench | — | 80.3% |
| LMCA | — | 29.3% |
| ForecastBench | — | 59.7 |
Math Qwen3 235B-A22B leads
MiniMax-M2.5: 26.9 (#253), Qwen3 235B-A22B: 50.4 (#57)
| Benchmark | MiniMax-M2.5 | Qwen3 235B-A22B |
|---|---|---|
| LMArena Math | 1378 | 1432 |
| OTIS Mock AIME 2024-2025 | — | 86.7% |
| ProofBench | 4% | — |
| Omni-MATH | — | 71.8% |
| MATH Level 5 | — | 68.9% |
| FrontierMath (Feb 2025 set) | — | 8.5% |
| FrontierMath Tier 4 (v1) | — | 0% |
Knowledge Qwen3 235B-A22B leads
MiniMax-M2.5: 39.2 (#135), Qwen3 235B-A22B: 49.6 (#73)
| Benchmark | MiniMax-M2.5 | Qwen3 235B-A22B |
|---|---|---|
| Vectara Hallucination Rate | 9.1% | 9.3% |
| LMArena Expert | 1379 | 1463 |
| GPQA Diamond | — | 80.1% |
| SimpleQA Verified | — | 40.4% |
| MMLU-Pro | — | 84.4% |
| Confabulations | — | 15.6% |
| GPQA (HELM) | — | 72.7% |
Multilingual Qwen3 235B-A22B leads
MiniMax-M2.5: 47.1 (#152), Qwen3 235B-A22B: 52.3 (#89)
| Benchmark | MiniMax-M2.5 | Qwen3 235B-A22B |
|---|---|---|
| LMArena Non-English | 1338 | 1409 |
| LMArena Chinese | 1393 | 1481 |
| LMArena French | 1362 | 1445 |
| LMArena German | 1362 | 1433 |
| LMArena Japanese | 1171 | 1399 |
| LMArena Korean | 1232 | 1391 |
| LMArena Russian | 1358 | 1411 |
| LMArena Spanish | 1354 | 1430 |
Instruction Following Qwen3 235B-A22B leads
MiniMax-M2.5: 71.5 (#148), Qwen3 235B-A22B: 72.6 (#136)
| Benchmark | MiniMax-M2.5 | Qwen3 235B-A22B |
|---|---|---|
| LMArena Instruction Following | 1353 | 1408 |
| IFEval | — | 83.5% |
Long Context Qwen3 235B-A22B leads
MiniMax-M2.5: 37.5 (#216), Qwen3 235B-A22B: 46.1 (#26)
| Benchmark | MiniMax-M2.5 | Qwen3 235B-A22B |
|---|---|---|
| LMArena Longer Query | 1366 | 1426 |
| Fiction.LiveBench | — | 75% |
| CL-bench | 11.4% | — |
| CL-bench Life | 6.3% | — |
Writing & Preference Qwen3 235B-A22B leads
MiniMax-M2.5: 53.9 (#153), Qwen3 235B-A22B: 59.6 (#108)
| Benchmark | MiniMax-M2.5 | Qwen3 235B-A22B |
|---|---|---|
| LMArena Text | 1359 | 1419 |
| LMArena Creative Writing | 1331 | 1384 |
| EQ-Bench Creative Writing | 1361 | 1366 |
| LMArena Multi-Turn | 1364 | 1432 |
| Short-Story Creative Writing | — | 83% |
| WildBench | — | 86.6% |
Frequently asked questions
Is MiniMax-M2.5 better than Qwen3 235B-A22B?
Qwen3 235B-A22B is the stronger model overall, scoring 43.5 to 38.3 on the Noometry Index. MiniMax-M2.5 costs 2.3× less per token, which makes it the better buy when Qwen3 235B-A22B's lead doesn't matter for your workload.
Which is cheaper, MiniMax-M2.5 or Qwen3 235B-A22B?
MiniMax-M2.5 is cheaper. It lists at $0.30 per million input tokens and $1.20 per million output tokens; Qwen3 235B-A22B lists at $0.70 and $2.80.
Is MiniMax-M2.5 or Qwen3 235B-A22B better for coding?
MiniMax-M2.5 scores higher on coding benchmarks: 48.1 versus 44.3 in the Noometry coding category.
Which has the bigger context window?
MiniMax-M2.5 does, with 205K tokens against 131K.
How many benchmarks do MiniMax-M2.5 and Qwen3 235B-A22B share?
24 benchmarks have published results for both models. MiniMax-M2.5 has 33 scored results on Noometry and Qwen3 235B-A22B has 49.