Model comparison
MiMo-V2.6-Pro vs Qwen2.5 72B Instruct
MiMo-V2.6-Pro is the stronger model overall, scoring 50.3 to 31.9 on the Noometry Index.
Last verified . 12 shared benchmarks.
Alibaba (Qwen)Summary
- They share 12 benchmarks with published results for both. MiMo-V2.6-Pro scores higher in 9 categories and Qwen2.5 72B Instruct in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where MiMo-V2.6-Pro leads 54.5 to 19.3.
- MiMo-V2.6-Pro is cheaper at $0.43 / $0.87 per million input/output tokens, against $1.40 / $5.60 for Qwen2.5 72B Instruct.
- MiMo-V2.6-Pro accepts more context: 1.05M tokens versus 131K.
Side by side
| MiMo-V2.6-Pro | Qwen2.5 72B Instruct | |
|---|---|---|
| Provider | Xiaomi | Alibaba (Qwen) |
| Noometry Index | 50.3 | 31.9 |
| Released | 2026-09-21 | 2024-09 |
| Weights | Open | Open |
| Context window | 1.05M | 131K |
| Max output | 131K | 8K |
| Input $ / M tokens | $0.43 | $1.40 |
| Output $ / M tokens | $0.87 | $5.60 |
| Results tracked | 19 | 43 |
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), Qwen2.5 72B Instruct: 33.2 (#260)
| Benchmark | MiMo-V2.6-Pro | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Coding | 1534 | 1292 |
| LMArena WebDev | 1629 | — |
| SciCode | 60.9% | — |
| WeirdML | — | 16% |
| BigCodeBench Instruct | — | 45.8% |
| BigCodeBench Complete | — | 55.9% |
| ALE-Bench | 1,158 | — |
Agentic & Tool Use MiMo-V2.6-Pro leads
MiMo-V2.6-Pro: 37.5 (#35), Qwen2.5 72B Instruct: 22.1 (#133)
| Benchmark | MiMo-V2.6-Pro | Qwen2.5 72B Instruct |
|---|---|---|
| APEX-Agents | 59.5% | — |
| TheAgentCompany | — | 5.7% |
| BALROG | — | 16.2% |
| METR Time Horizons | — | 35.8% |
Reasoning MiMo-V2.6-Pro leads
MiMo-V2.6-Pro: 43.1 (#50), Qwen2.5 72B Instruct: 22.3 (#199)
| Benchmark | MiMo-V2.6-Pro | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Hard Prompts | 1512 | 1271 |
| CritPt | 26.6% | — |
| DTBench | — | 62.9% |
| LMCA | — | 13.4% |
| BIG-Bench Hard | — | 79.8% |
| Epoch Capabilities Index | — | 129 |
| ForecastBench | — | 57.5 |
| HellaSwag | — | 84.8% |
| PIQA | — | 82.6% |
| WinoGrande | — | 82.3% |
Math MiMo-V2.6-Pro leads
MiMo-V2.6-Pro: 54.5 (#45), Qwen2.5 72B Instruct: 19.3 (#287)
| Benchmark | MiMo-V2.6-Pro | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Math | 1494 | 1283 |
| OTIS Mock AIME 2024-2025 | — | 8.1% |
| ProofBench | 70% | — |
| Omni-MATH | — | 33% |
| MATH Level 5 | — | 63.2% |
Knowledge MiMo-V2.6-Pro leads
MiMo-V2.6-Pro: 43.5 (#92), Qwen2.5 72B Instruct: 27.0 (#253)
| Benchmark | MiMo-V2.6-Pro | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Expert | 1543 | 1245 |
| GPQA Diamond | — | 49.1% |
| MMLU-Pro | — | 63.1% |
| Confabulations | — | 19.1% |
| GPQA (HELM) | — | 42.6% |
| ARC (AI2) Challenge | — | 94.5% |
| MMLU | — | 85.3% |
| TriviaQA | — | 71.9% |
Multimodal Not comparable
MiMo-V2.6-Pro: 40.8 (#43), Qwen2.5 72B Instruct: —
| Benchmark | MiMo-V2.6-Pro | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Vision | 1264 | — |
Multilingual MiMo-V2.6-Pro leads
MiMo-V2.6-Pro: 56.9 (#14), Qwen2.5 72B Instruct: 41.0 (#213)
| Benchmark | MiMo-V2.6-Pro | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Non-English | 1474 | 1252 |
| LMArena Chinese | 1529 | 1272 |
| LMArena Russian | 1480 | 1264 |
| LMArena French | — | 1280 |
| LMArena German | — | 1234 |
| LMArena Japanese | — | 1180 |
| LMArena Korean | — | 1188 |
| LMArena Spanish | — | 1256 |
Instruction Following MiMo-V2.6-Pro leads
MiMo-V2.6-Pro: 78.2 (#12), Qwen2.5 72B Instruct: 65.5 (#221)
| Benchmark | MiMo-V2.6-Pro | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Instruction Following | 1493 | 1254 |
| IFEval | — | 80.6% |
Long Context MiMo-V2.6-Pro leads
MiMo-V2.6-Pro: 46.0 (#27), Qwen2.5 72B Instruct: 38.9 (#188)
| Benchmark | MiMo-V2.6-Pro | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Longer Query | 1501 | 1282 |
Writing & Preference MiMo-V2.6-Pro leads
MiMo-V2.6-Pro: 66.8 (#33), Qwen2.5 72B Instruct: 46.7 (#215)
| Benchmark | MiMo-V2.6-Pro | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Text | 1492 | 1269 |
| LMArena Creative Writing | 1468 | 1221 |
| LMArena Multi-Turn | 1464 | 1272 |
| WildBench | — | 80.2% |
Frequently asked questions
Is MiMo-V2.6-Pro better than Qwen2.5 72B Instruct?
MiMo-V2.6-Pro is the stronger model overall, scoring 50.3 to 31.9 on the Noometry Index.
Which is cheaper, MiMo-V2.6-Pro or Qwen2.5 72B Instruct?
MiMo-V2.6-Pro is cheaper. It lists at $0.43 per million input tokens and $0.87 per million output tokens; Qwen2.5 72B Instruct lists at $1.40 and $5.60.
Is MiMo-V2.6-Pro or Qwen2.5 72B Instruct better for coding?
MiMo-V2.6-Pro scores higher on coding benchmarks: 55.5 versus 33.2 in the Noometry coding category.
Which has the bigger context window?
MiMo-V2.6-Pro does, with 1.05M tokens against 131K.
How many benchmarks do MiMo-V2.6-Pro and Qwen2.5 72B Instruct share?
12 benchmarks have published results for both models. MiMo-V2.6-Pro has 19 scored results on Noometry and Qwen2.5 72B Instruct has 43.