Model comparison
MiniMax-M3 vs Qwen3 8B
MiniMax-M3 is the stronger model overall, scoring 43.8 to 33.7 on the Noometry Index. Qwen3 8B costs 1.7× less per token, which makes it the better buy when MiniMax-M3's lead doesn't matter for your workload.
Last verified . 8 shared benchmarks.
Summary
- They share 8 benchmarks with published results for both. MiniMax-M3 scores higher in 5 categories and Qwen3 8B in 1 category; 6 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where MiniMax-M3 leads 58.4 to 36.1.
- The biggest single-benchmark swing is GPQA Diamond: 90.9% for MiniMax-M3 and 56.8% for Qwen3 8B.
- Qwen3 8B is cheaper at $0.18 / $0.70 per million input/output tokens, against $0.30 / $1.20 for MiniMax-M3.
- MiniMax-M3 accepts more context: 1M tokens versus 131K.
Side by side
| MiniMax-M3 | Qwen3 8B | |
|---|---|---|
| Provider | MiniMax | Alibaba (Qwen) |
| Noometry Index | 43.8 | 33.7 |
| Released | 2026-06-01 | 2025-04 |
| Weights | Open | Open |
| Context window | 1M | 131K |
| Max output | 512K | 8K |
| Input $ / M tokens | $0.30 | $0.18 |
| Output $ / M tokens | $1.20 | $0.70 |
| Results tracked | 41 | 11 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiniMax-M3 leads
MiniMax-M3: 41.8 (#118), Qwen3 8B: 34.0 (#248)
| Benchmark | MiniMax-M3 | Qwen3 8B |
|---|---|---|
| SciCode | 47.1% | 22.6% |
| FrontierCode | 14.7% | — |
| LMArena WebDev | 1482 | — |
| LMArena Coding | 1469 | — |
| ALE-Bench | 640.02 | — |
Agentic & Tool Use Qwen3 8B leads
MiniMax-M3: 22.6 (#130), Qwen3 8B: 30.2 (#78)
| Benchmark | MiniMax-M3 | Qwen3 8B |
|---|---|---|
| APEX-Agents | 37.7% | — |
| Berkeley Function Calling Leaderboard | — | 42.6% |
| OSWorld 2.0 | 4.6% | — |
| GBAEval | 0.9% | — |
| Vending-Bench 2 | 2,158 | — |
Reasoning MiniMax-M3 leads
MiniMax-M3: 30.1 (#87), Qwen3 8B: 16.6 (#303)
| Benchmark | MiniMax-M3 | Qwen3 8B |
|---|---|---|
| CritPt | 3.7% | 0% |
| Chess Puzzles | 14% | 5% |
| DTBench | 78.9% | 59.7% |
| LMCA | 33.7% | 8.8% |
| Epoch Capabilities Index | 146.95 | 136.17 |
| SimpleBench | 45.8% | — |
| NYT Connections (extended) | 65.1% | — |
| LMArena Hard Prompts | 1447 | — |
| Mystery Game Puzzles | 8% | — |
| Surface Evolver Bench | 55% | — |
| ForecastBench | 61.4 | — |
Math MiniMax-M3 leads
MiniMax-M3: 40.0 (#95), Qwen3 8B: 34.9 (#191)
| Benchmark | MiniMax-M3 | Qwen3 8B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 71.1% | 56.1% |
| ProofBench | 18% | — |
| LMArena Math | 1429 | — |
Knowledge MiniMax-M3 leads
MiniMax-M3: 58.4 (#35), Qwen3 8B: 36.1 (#173)
| Benchmark | MiniMax-M3 | Qwen3 8B |
|---|---|---|
| GPQA Diamond | 90.9% | 56.8% |
| Vectara Hallucination Rate | — | 4.8% |
| LMArena Expert | 1461 | — |
Multimodal Not comparable
MiniMax-M3: 40.2 (#51), Qwen3 8B: —
| Benchmark | MiniMax-M3 | Qwen3 8B |
|---|---|---|
| LMArena Vision | 1253 | — |
| LMArena Document | 1435 | — |
Multilingual Not comparable
MiniMax-M3: 53.0 (#75), Qwen3 8B: —
| Benchmark | MiniMax-M3 | Qwen3 8B |
|---|---|---|
| LMArena Non-English | 1420 | — |
| LMArena Chinese | 1463 | — |
| LMArena French | 1447 | — |
| LMArena German | 1426 | — |
| LMArena Japanese | 1381 | — |
| LMArena Korean | 1372 | — |
| LMArena Russian | 1428 | — |
| LMArena Spanish | 1432 | — |
Instruction Following Not comparable
MiniMax-M3: 75.5 (#62), Qwen3 8B: —
| Benchmark | MiniMax-M3 | Qwen3 8B |
|---|---|---|
| LMArena Instruction Following | 1433 | — |
Long Context MiniMax-M3 leads
MiniMax-M3: 44.2 (#72), Qwen3 8B: 37.9 (#210)
| Benchmark | MiniMax-M3 | Qwen3 8B |
|---|---|---|
| Fiction.LiveBench | — | 62.1% |
| LMArena Longer Query | 1445 | — |
Writing & Preference Not comparable
MiniMax-M3: 62.1 (#83), Qwen3 8B: —
| Benchmark | MiniMax-M3 | Qwen3 8B |
|---|---|---|
| LMArena Text | 1433 | — |
| LMArena Creative Writing | 1404 | — |
| EQ-Bench 4 | 1150 | — |
| LMArena Multi-Turn | 1442 | — |
Frequently asked questions
Is MiniMax-M3 better than Qwen3 8B?
MiniMax-M3 is the stronger model overall, scoring 43.8 to 33.7 on the Noometry Index. Qwen3 8B costs 1.7× less per token, which makes it the better buy when MiniMax-M3's lead doesn't matter for your workload.
Which is cheaper, MiniMax-M3 or Qwen3 8B?
Qwen3 8B is cheaper. It lists at $0.18 per million input tokens and $0.70 per million output tokens; MiniMax-M3 lists at $0.30 and $1.20.
Is MiniMax-M3 or Qwen3 8B better for coding?
MiniMax-M3 scores higher on coding benchmarks: 41.8 versus 34.0 in the Noometry coding category.
Which has the bigger context window?
MiniMax-M3 does, with 1M tokens against 131K.
How many benchmarks do MiniMax-M3 and Qwen3 8B share?
8 benchmarks have published results for both models. MiniMax-M3 has 41 scored results on Noometry and Qwen3 8B has 11.