Model comparison
MiMo-V2.6-Pro vs Qwen3.8 Max
Qwen3.8 Max is the stronger model overall, scoring 56.8 to 50.3 on the Noometry Index. MiMo-V2.6-Pro costs 5.5× less per token, which makes it the better buy when Qwen3.8 Max's lead doesn't matter for your workload.
Last verified . 18 shared benchmarks.
Alibaba (Qwen)Summary
- They share 18 benchmarks with published results for both. MiMo-V2.6-Pro scores higher in 5 categories and Qwen3.8 Max in 5 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3.8 Max leads 73.2 to 54.5.
- The biggest single-benchmark swing is ProofBench: 70% for MiMo-V2.6-Pro and 58% for Qwen3.8 Max.
- MiMo-V2.6-Pro is cheaper at $0.43 / $0.87 per million input/output tokens, against $2 / $6 for Qwen3.8 Max.
- MiMo-V2.6-Pro accepts more context: 1.05M tokens versus 1M.
- MiMo-V2.6-Pro has downloadable open weights; the other is API-only.
Side by side
| MiMo-V2.6-Pro | Qwen3.8 Max | |
|---|---|---|
| Provider | Xiaomi | Alibaba (Qwen) |
| Noometry Index | 50.3 | 56.8 |
| Released | 2026-09-21 | 2026-08-02 |
| Weights | Open | Proprietary |
| Context window | 1.05M | 1M |
| Max output | 131K | 131K |
| Input $ / M tokens | $0.43 | $2 |
| Output $ / M tokens | $0.87 | $6 |
| Results tracked | 19 | 39 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiMo-V2.6-Pro leads
MiMo-V2.6-Pro: 55.5 (#23), Qwen3.8 Max: 53.5 (#29)
| Benchmark | MiMo-V2.6-Pro | Qwen3.8 Max |
|---|---|---|
| LMArena WebDev | 1629 | 1674 |
| SciCode | 60.9% | 53.2% |
| LMArena Coding | 1534 | 1502 |
| DeepSWE | — | 57.5% |
| FrontierSWE | — | 17.8% |
| ALE-Bench | 1,158 | — |
Agentic & Tool Use Qwen3.8 Max leads
MiMo-V2.6-Pro: 37.5 (#35), Qwen3.8 Max: 45.4 (#14)
| Benchmark | MiMo-V2.6-Pro | Qwen3.8 Max |
|---|---|---|
| APEX-Agents | 59.5% | 63.3% |
| τ²-bench Banking | — | 55.1% |
| GDP.pdf | — | 23.2% |
Reasoning Qwen3.8 Max leads
MiMo-V2.6-Pro: 43.1 (#50), Qwen3.8 Max: 54.4 (#26)
| Benchmark | MiMo-V2.6-Pro | Qwen3.8 Max |
|---|---|---|
| CritPt | 26.6% | 20% |
| LMArena Hard Prompts | 1512 | 1496 |
| NYT Connections (extended) | — | 88.3% |
| Chess Puzzles | — | 40% |
| Mystery Game Puzzles | — | 38% |
| DTBench | — | 92% |
| LMCA | — | 46.2% |
| Epoch Capabilities Index | — | 156.41 |
Math Qwen3.8 Max leads
MiMo-V2.6-Pro: 54.5 (#45), Qwen3.8 Max: 73.2 (#20)
| Benchmark | MiMo-V2.6-Pro | Qwen3.8 Max |
|---|---|---|
| ProofBench | 70% | 58% |
| LMArena Math | 1494 | 1499 |
| FrontierMath (Tiers 1-3) | — | 74.7% |
| FrontierMath Tier 4 | — | 46.3% |
| OTIS Mock AIME 2024-2025 | — | 100% |
Knowledge Qwen3.8 Max leads
MiMo-V2.6-Pro: 43.5 (#92), Qwen3.8 Max: 61.7 (#27)
| Benchmark | MiMo-V2.6-Pro | Qwen3.8 Max |
|---|---|---|
| LMArena Expert | 1543 | 1507 |
| GPQA Diamond | — | 92.7% |
| SimpleQA Verified | — | 47.3% |
Multimodal MiMo-V2.6-Pro leads
MiMo-V2.6-Pro: 40.8 (#43), Qwen3.8 Max: 37.2 (#75)
| Benchmark | MiMo-V2.6-Pro | Qwen3.8 Max |
|---|---|---|
| LMArena Vision | 1264 | 1314 |
| Furniture Assembly | — | 20% |
Multilingual Too close to call
MiMo-V2.6-Pro: 56.9 (#14), Qwen3.8 Max: 56.7 (#18)
| Benchmark | MiMo-V2.6-Pro | Qwen3.8 Max |
|---|---|---|
| LMArena Non-English | 1474 | 1472 |
| LMArena Chinese | 1529 | 1538 |
| LMArena Russian | 1480 | 1481 |
| LMArena French | — | 1503 |
| LMArena German | — | 1483 |
| LMArena Japanese | — | 1467 |
| LMArena Korean | — | 1461 |
| LMArena Spanish | — | 1492 |
Instruction Following Too close to call
MiMo-V2.6-Pro: 78.2 (#12), Qwen3.8 Max: 77.6 (#17)
| Benchmark | MiMo-V2.6-Pro | Qwen3.8 Max |
|---|---|---|
| LMArena Instruction Following | 1493 | 1479 |
Long Context Too close to call
MiMo-V2.6-Pro: 46.0 (#27), Qwen3.8 Max: 45.6 (#31)
| Benchmark | MiMo-V2.6-Pro | Qwen3.8 Max |
|---|---|---|
| LMArena Longer Query | 1501 | 1489 |
Writing & Preference Too close to call
MiMo-V2.6-Pro: 66.8 (#33), Qwen3.8 Max: 67.1 (#30)
| Benchmark | MiMo-V2.6-Pro | Qwen3.8 Max |
|---|---|---|
| LMArena Text | 1492 | 1483 |
| LMArena Creative Writing | 1468 | 1479 |
| LMArena Multi-Turn | 1464 | 1489 |
Frequently asked questions
Is MiMo-V2.6-Pro better than Qwen3.8 Max?
Qwen3.8 Max is the stronger model overall, scoring 56.8 to 50.3 on the Noometry Index. MiMo-V2.6-Pro costs 5.5× less per token, which makes it the better buy when Qwen3.8 Max's lead doesn't matter for your workload.
Which is cheaper, MiMo-V2.6-Pro or Qwen3.8 Max?
MiMo-V2.6-Pro is cheaper. It lists at $0.43 per million input tokens and $0.87 per million output tokens; Qwen3.8 Max lists at $2 and $6.
Is MiMo-V2.6-Pro or Qwen3.8 Max better for coding?
MiMo-V2.6-Pro scores higher on coding benchmarks: 55.5 versus 53.5 in the Noometry coding category.
Which has the bigger context window?
MiMo-V2.6-Pro does, with 1.05M tokens against 1M.
How many benchmarks do MiMo-V2.6-Pro and Qwen3.8 Max share?
18 benchmarks have published results for both models. MiMo-V2.6-Pro has 19 scored results on Noometry and Qwen3.8 Max has 39.