Model comparison
DeepSeek-V3 vs MiMo-V2.5-Pro
MiMo-V2.5-Pro is the stronger model overall, scoring 44.8 to 39.2 on the Noometry Index.
Last verified . 22 shared benchmarks.
DeepSeekSummary
- They share 22 benchmarks with published results for both. DeepSeek-V3 scores higher in 0 categories and MiMo-V2.5-Pro in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in long context, where MiMo-V2.5-Pro leads 45.4 to 34.0.
- The biggest single-benchmark swing is DTBench: 64.8% for DeepSeek-V3 and 84.5% for MiMo-V2.5-Pro.
- DeepSeek-V3 is cheaper at $0.24 / $0.90 per million input/output tokens, against $0.43 / $0.87 for MiMo-V2.5-Pro.
- MiMo-V2.5-Pro accepts more context: 1.05M tokens versus 164K.
Side by side
| DeepSeek-V3 | MiMo-V2.5-Pro | |
|---|---|---|
| Provider | DeepSeek | Xiaomi |
| Noometry Index | 39.2 | 44.8 |
| Released | 2024-12-26 | 2026-04-22 |
| Weights | Open | Open |
| Context window | 164K | 1.05M |
| Max output | 164K | 131K |
| Input $ / M tokens | $0.24 | $0.43 |
| Output $ / M tokens | $0.90 | $0.87 |
| Results tracked | 60 | 27 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiMo-V2.5-Pro leads
DeepSeek-V3: 41.9 (#99), MiMo-V2.5-Pro: 47.0 (#58)
| Benchmark | DeepSeek-V3 | MiMo-V2.5-Pro |
|---|---|---|
| SciCode | 35.8% | 50.2% |
| LMArena Coding | 1368 | 1502 |
| Aider Polyglot | 55.1% | — |
| LMArena WebDev | — | 1477 |
| WeirdML | 36.1% | — |
| BigCodeBench Instruct | 50% | — |
| LiveBench Coding | 70.9% | — |
| BigCodeBench Complete | 62.2% | — |
| ALE-Bench | — | 899.8 |
| HumanEval+ | 86.6% | — |
| MBPP+ | 73% | — |
Agentic & Tool Use Not comparable
DeepSeek-V3: —, MiMo-V2.5-Pro: —
| Benchmark | DeepSeek-V3 | MiMo-V2.5-Pro |
|---|---|---|
| METR Time Horizons | 49.6% | — |
Reasoning MiMo-V2.5-Pro leads
DeepSeek-V3: 20.0 (#226), MiMo-V2.5-Pro: 25.9 (#124)
| Benchmark | DeepSeek-V3 | MiMo-V2.5-Pro |
|---|---|---|
| CritPt | 0% | 4% |
| LMArena Hard Prompts | 1365 | 1487 |
| DTBench | 64.8% | 84.5% |
| LMCA | 15.5% | 29.5% |
| SimpleBench | 27.2% | — |
| Kagi LLM Benchmark | 52.3% | — |
| NYT Connections (extended) | — | 34.4% |
| LiveBench Reasoning | 65.8% | — |
| LiveBench Data Analysis | 60.9% | — |
| BIG-Bench Hard | 87.5% | — |
| Epoch Capabilities Index | 135.95 | — |
| ForecastBench | 59.1 | — |
| HellaSwag | 88.9% | — |
| LiveBench | 66.9% | — |
| PIQA | 84.7% | — |
| WinoGrande | 85.2% | — |
Math MiMo-V2.5-Pro leads
DeepSeek-V3: 30.4 (#208), MiMo-V2.5-Pro: 38.6 (#84)
| Benchmark | DeepSeek-V3 | MiMo-V2.5-Pro |
|---|---|---|
| LMArena Math | 1373 | 1480 |
| OTIS Mock AIME 2024-2025 | 37.8% | — |
| ProofBench | — | 22% |
| Omni-MATH | 40.3% | — |
| LiveBench Math | 73.5% | — |
| MATH Level 5 | 75.5% | — |
| FrontierMath (Feb 2025 set) | 1.7% | — |
Knowledge MiMo-V2.5-Pro leads
DeepSeek-V3: 37.5 (#147), MiMo-V2.5-Pro: 41.5 (#100)
| Benchmark | DeepSeek-V3 | MiMo-V2.5-Pro |
|---|---|---|
| LMArena Expert | 1352 | 1501 |
| GPQA Diamond | 67.6% | — |
| MMLU-Pro | 72.3% | — |
| Confabulations | 26.1% | — |
| Vectara Hallucination Rate | 6.1% | — |
| GPQA (HELM) | 53.8% | — |
| ARC (AI2) Challenge | 95.3% | — |
| MMLU | 87.2% | — |
| TriviaQA | 82.9% | — |
Multilingual MiMo-V2.5-Pro leads
DeepSeek-V3: 48.8 (#138), MiMo-V2.5-Pro: 55.3 (#33)
| Benchmark | DeepSeek-V3 | MiMo-V2.5-Pro |
|---|---|---|
| LMArena Non-English | 1358 | 1449 |
| LMArena Chinese | 1390 | 1507 |
| LMArena French | 1386 | 1489 |
| LMArena German | 1374 | 1455 |
| LMArena Japanese | 1333 | 1418 |
| LMArena Korean | 1321 | 1436 |
| LMArena Russian | 1372 | 1450 |
| LMArena Spanish | 1358 | 1470 |
Instruction Following MiMo-V2.5-Pro leads
DeepSeek-V3: 73.0 (#125), MiMo-V2.5-Pro: 77.6 (#21)
| Benchmark | DeepSeek-V3 | MiMo-V2.5-Pro |
|---|---|---|
| LMArena Instruction Following | 1345 | 1477 |
| LiveBench Instruction Following | 81.5% | — |
| IFEval | 83.2% | — |
Long Context MiMo-V2.5-Pro leads
DeepSeek-V3: 34.0 (#248), MiMo-V2.5-Pro: 45.4 (#38)
| Benchmark | DeepSeek-V3 | MiMo-V2.5-Pro |
|---|---|---|
| LMArena Longer Query | 1353 | 1482 |
| Fiction.LiveBench | 50% | — |
Writing & Preference MiMo-V2.5-Pro leads
DeepSeek-V3: 57.6 (#124), MiMo-V2.5-Pro: 65.5 (#47)
| Benchmark | DeepSeek-V3 | MiMo-V2.5-Pro |
|---|---|---|
| LMArena Text | 1375 | 1465 |
| LMArena Creative Writing | 1364 | 1440 |
| EQ-Bench Creative Writing | 1472 | 1493 |
| LMArena Multi-Turn | 1389 | 1477 |
| Short-Story Creative Writing | 77% | — |
| WildBench | 83% | — |
| EQ-Bench 4 | — | 1208 |
| LiveBench Language | 49.1% | — |
Frequently asked questions
Is DeepSeek-V3 better than MiMo-V2.5-Pro?
MiMo-V2.5-Pro is the stronger model overall, scoring 44.8 to 39.2 on the Noometry Index.
Which is cheaper, DeepSeek-V3 or MiMo-V2.5-Pro?
DeepSeek-V3 is cheaper. It lists at $0.24 per million input tokens and $0.90 per million output tokens; MiMo-V2.5-Pro lists at $0.43 and $0.87.
Is DeepSeek-V3 or MiMo-V2.5-Pro better for coding?
MiMo-V2.5-Pro scores higher on coding benchmarks: 47.0 versus 41.9 in the Noometry coding category.
Which has the bigger context window?
MiMo-V2.5-Pro does, with 1.05M tokens against 164K.
How many benchmarks do DeepSeek-V3 and MiMo-V2.5-Pro share?
22 benchmarks have published results for both models. DeepSeek-V3 has 60 scored results on Noometry and MiMo-V2.5-Pro has 27.