Qwen3.8 27B vs MiMo-V2.5-Pro
Updated October 21, 2026. 56 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.
Decision reading
Pick Qwen3.8 27B for coding: it scores 73.6 on the Vibe Coding Index against 40.0 for MiMo-V2.5-Pro.
Pick MiMo-V2.5-Pro when output tokens cost 2.2x less, or when a solved task costs $0.03 against $1.26.
- Shared results
- 56
- Qwen3.8 27B only
- 29
- MiMo-V2.5-Pro only
- 14
- Like-for-like categories
- 7 of 9
Shared results can support a head-to-head reading. Results only one model has describe coverage, not superiority.
Which one for your work
Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.
Coding
Vibe Coding Index, four public coding benchmarks
Qwen3.8 27B
Qwen3.8 27B scores 73.6 against 40.0.
2 and 1 benchmarks behind the scores
Intelligence
AA Intelligence Index
Qwen3.8 27B
Qwen3.8 27B leads, 33.7 against 26.0.
Same board, both measured
Agentic work
AA Agentic Index
Qwen3.8 27B
Qwen3.8 27B leads, 45.8 against 21.3.
Same board, both measured
Reasoning
GPQA Diamond
Qwen3.8 27B
Qwen3.8 27B leads, 88.9% against 82.6%.
Same board, both measured
Knowledge
MMLU Pro
MiMo-V2.5-Pro
MiMo-V2.5-Pro leads, 84.6% against 84.3%.
Same board, both measured
Web design
Design Arena: websites
No call
Only MiMo-V2.5-Pro has a Design Arena: websites result, 1267.
Directional only
Long documents
Listed context window
MiMo-V2.5-Pro
MiMo-V2.5-Pro takes 1.05M tokens against 1M.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
MiMo-V2.5-Pro
Qwen3.8 27B $0.0007, MiMo-V2.5-Pro $0.0006 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
MiMo-V2.5-Pro
Qwen3.8 27B $1.26, MiMo-V2.5-Pro $0.03.
Published attempt cost and pass rate
At a glance
- Coding index
- Qwen3.8 27B 73.6 from 2 benchmarks, MiMo-V2.5-Pro 40.0 from 1 benchmark.
- Widest category
- Agentic work, 15.1 points apart on average over 5 shared benchmarks.
- Widest single gap
- Vibe Code Bench (Vals AI): Qwen3.8 27B 64.8% against MiMo-V2.5-Pro 34.1%, 30.7 points apart on a 0 to 100 scale.
- Output price
- Qwen3.8 27B $1.35 against MiMo-V2.5-Pro $0.609 per million output tokens, 2.2x apart.
- Context window
- Qwen3.8 27B takes 1M tokens, MiMo-V2.5-Pro 1.05M.
Benchmark by benchmark
First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.
| Category | Qwen3.8 27B | Bars | MiMo-V2.5-Pro | Ties |
|---|---|---|---|---|
| Coding | 8 | 4 | ||
| Intelligence | 1 | 0 | ||
| Agentic work | 5 | 0 | ||
| Reasoning | 2 | 4 | ||
| Math | 0 | 3 | ||
| Knowledge | 1 | 7 | ||
| General | 4 | 17 | ||
| All shared boards | 21 | 35 |
CodingLike-for-likeFour of these boards feed the Vibe Coding Index84
| Benchmark | Qwen3.8 27B | MiMo-V2.5-Pro | Leader |
|---|---|---|---|
| LiveBench coding | 86.8% | Not measured | Only Qwen3.8 27B measured |
| LMArena WebDev | 1593 | 1480 | Qwen3.8 27B |
| AA Coding Index | 68.1 | 60.2 | Qwen3.8 27B |
| LiveCodeBench (Vals AI) | 84.0% | 81.3% | Qwen3.8 27B |
| SWE-bench Verified (Vals AI) | 86.0% | 74.0% | Qwen3.8 27B |
| Terminal-Bench 4 (Vals AI) | Not measured | 0.5% | Only MiMo-V2.5-Pro measured |
| Vibe Code Bench (Vals AI) | 64.8% | 34.1% | Qwen3.8 27B |
| Code migration (Vals AI) | 14.2% | 21.6% | MiMo-V2.5-Pro |
| ProgramBench (Vals AI) | 0.0% | Not measured | Only Qwen3.8 27B measured |
| IOI (Vals AI) | 39.1% | Not measured | Only Qwen3.8 27B measured |
| Terminal-Bench 2.1 (Vals AI) | 58.4% | 57.3% | Qwen3.8 27B |
| SkillsBench (Vals AI) | 38.1% | Not measured | Only Qwen3.8 27B measured |
| CyberBench (Vals AI) | 67.8% | Not measured | Only Qwen3.8 27B measured |
| SciCode | 46.6% | 50.6% | MiMo-V2.5-Pro |
| Terminal-Bench Hard (Artificial Analysis run) | Not measured | 43.2% | Only MiMo-V2.5-Pro measured |
| ALE-Bench | Not measured | 899.8 | Only MiMo-V2.5-Pro measured |
| LMArena Text: coding | 1483 | 1501 | MiMo-V2.5-Pro |
| LMArena Text: software and IT | 1478 | 1495 | MiMo-V2.5-Pro |
| LMArena WebDev: React | 1602 | 1469 | Qwen3.8 27B |
| LMArena WebDev: HTML | 1570 | 1473 | Qwen3.8 27B |
| LMArena WebDev: from an image | 1577 | Not measured | Only Qwen3.8 27B measured |
| Neither model has a result on the other 17 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | Qwen3.8 27B | MiMo-V2.5-Pro | Leader |
|---|---|---|---|
| AA Intelligence Index | 33.7 | 26.0 | Qwen3.8 27B |
| Epoch Capabilities Index | 151.7 | Not measured | Only Qwen3.8 27B measured |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index50
| Benchmark | Qwen3.8 27B | MiMo-V2.5-Pro | Leader |
|---|---|---|---|
| AA Agentic Index | 45.8 | 21.3 | Qwen3.8 27B |
| LiveBench research | 59.4% | Not measured | Only Qwen3.8 27B measured |
| LiveBench browser | 32.0% | Not measured | Only Qwen3.8 27B measured |
| LiveBench long agentic | 27.9% | Not measured | Only Qwen3.8 27B measured |
| Finance Agent (Vals AI) | 48.5% | 41.5% | Qwen3.8 27B |
| Tax Agent Bench (Vals AI) | 30.9% | 11.5% | Qwen3.8 27B |
| Legal Research Bench (Vals AI) | 36.1% | 15.9% | Qwen3.8 27B |
| Excel Modeling Benchmark (Vals AI) | 59.7% | 55.2% | Qwen3.8 27B |
| Terminal-Bench-Science (Vals AI) | 0.0% | Not measured | Only Qwen3.8 27B measured |
| τ²-Bench Telecom | Not measured | 94.2% | Only MiMo-V2.5-Pro measured |
| APEX-Agents | 47.5% | Not measured | Only Qwen3.8 27B measured |
| Neither model has a result on the other 28 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index24
| Benchmark | Qwen3.8 27B | MiMo-V2.5-Pro | Leader |
|---|---|---|---|
| GPQA Diamond | 88.9% | 82.6% | Qwen3.8 27B |
| LiveBench reasoning | 80.0% | Not measured | Only Qwen3.8 27B measured |
| MMMU Pro (Vals AI) | 83.9% | Not measured | Only Qwen3.8 27B measured |
| GPQA Diamond (Artificial Analysis run) | 90.5% | 86.6% | Qwen3.8 27B |
| Humanity's Last Exam (Artificial Analysis run) | 33.9% | 35.7% | MiMo-V2.5-Pro |
| ARC-AGI | 87.5% | Not measured | Only Qwen3.8 27B measured |
| ARC-AGI-2 | 42.4% | Not measured | Only Qwen3.8 27B measured |
| LMArena Text: hard prompts | 1459 | 1486 | MiMo-V2.5-Pro |
| LMArena Text: expert | 1479 | 1501 | MiMo-V2.5-Pro |
| LMArena Text: hard prompts in English | 1461 | 1494 | MiMo-V2.5-Pro |
| LMArena Vision: diagrams | 1285 | Not measured | Only Qwen3.8 27B measured |
| LMArena Vision: homework | 1295 | Not measured | Only Qwen3.8 27B measured |
| Neither model has a result on the other 9 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index03
| Benchmark | Qwen3.8 27B | MiMo-V2.5-Pro | Leader |
|---|---|---|---|
| ProofBench | 16.0% | 22.0% | MiMo-V2.5-Pro |
| LMArena Text: math | 1455 | 1481 | MiMo-V2.5-Pro |
| LMArena Text: mathematical work | 1458 | 1491 | MiMo-V2.5-Pro |
| Neither model has a result on the other 10 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index17
| Benchmark | Qwen3.8 27B | MiMo-V2.5-Pro | Leader |
|---|---|---|---|
| MMLU Pro | 84.3% | 84.6% | MiMo-V2.5-Pro |
| LegalBench (Vals AI) | 82.4% | 77.1% | Qwen3.8 27B |
| CorpFin (Vals AI) | Not measured | 61.4% | Only MiMo-V2.5-Pro measured |
| TaxEval (Vals AI) | 70.8% | 73.8% | MiMo-V2.5-Pro |
| MortgageTax (Vals AI) | 64.9% | Not measured | Only Qwen3.8 27B measured |
| MedCode (Vals AI) | 28.7% | 32.5% | MiMo-V2.5-Pro |
| LMArena Text: science | 1467 | 1482 | MiMo-V2.5-Pro |
| LMArena Text: medicine and healthcare | 1437 | 1473 | MiMo-V2.5-Pro |
| LMArena Text: legal and government | 1447 | 1474 | MiMo-V2.5-Pro |
| LMArena Text: business and finance | 1438 | 1466 | MiMo-V2.5-Pro |
| Neither model has a result on the other 7 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index9 boards, none shared
| Benchmark | Qwen3.8 27B | MiMo-V2.5-Pro | Leader |
|---|---|---|---|
| Design Arena: websites | Not measured | 1267 | Only MiMo-V2.5-Pro measured |
| Design Arena: UI components | Not measured | 1248 | Only MiMo-V2.5-Pro measured |
| Design Arena: game dev | Not measured | 1261 | Only MiMo-V2.5-Pro measured |
| Design Arena: data viz | Not measured | 1257 | Only MiMo-V2.5-Pro measured |
| Design Arena: 3D | Not measured | 1244 | Only MiMo-V2.5-Pro measured |
| Design Arena: all code categories | Not measured | 1266 | Only MiMo-V2.5-Pro measured |
| Design Arena: SVG | Not measured | 1181 | Only MiMo-V2.5-Pro measured |
| Design Arena: ASCII art | Not measured | 1157 | Only MiMo-V2.5-Pro measured |
| LiveBench design | 75.9% | Not measured | Only Qwen3.8 27B measured |
GeneralLike-for-likePublic boards, shown as published, kept out of the index417
| Benchmark | Qwen3.8 27B | MiMo-V2.5-Pro | Leader |
|---|---|---|---|
| LiveBench data analysis | 76.6% | Not measured | Only Qwen3.8 27B measured |
| SAGE (Vals AI) | 52.4% | Not measured | Only Qwen3.8 27B measured |
| MedScribe (Vals AI) | 83.8% | 83.7% | Qwen3.8 27B |
| IFBench | Not measured | 79.9% | Only MiMo-V2.5-Pro measured |
| Long context reasoning (AA-LCR) | 82.0% | 79.7% | Qwen3.8 27B |
| DTBench | 88.0% | 84.5% | Qwen3.8 27B |
| LMCA | 41.4% | 29.5% | Qwen3.8 27B |
| Surface Evolver Bench | 45.0% | Not measured | Only Qwen3.8 27B measured |
| LMArena Text | 1440 | 1464 | MiMo-V2.5-Pro |
| LMArena Text: instruction following | 1436 | 1475 | MiMo-V2.5-Pro |
| LMArena Text: creative writing | 1383 | 1438 | MiMo-V2.5-Pro |
| LMArena Vision | 1269 | Not measured | Only Qwen3.8 27B measured |
| LMArena Text: English | 1448 | 1477 | MiMo-V2.5-Pro |
| LMArena Text: non-English | 1428 | 1448 | MiMo-V2.5-Pro |
| LMArena Text: Chinese | 1501 | 1507 | MiMo-V2.5-Pro |
| LMArena Text: German | 1443 | 1459 | MiMo-V2.5-Pro |
| LMArena Text: Japanese | 1387 | 1410 | MiMo-V2.5-Pro |
| LMArena Text: Korean | 1384 | 1435 | MiMo-V2.5-Pro |
| LMArena Text: French | 1465 | 1487 | MiMo-V2.5-Pro |
| LMArena Text: Spanish | 1449 | 1469 | MiMo-V2.5-Pro |
| LMArena Text: Russian | 1415 | 1449 | MiMo-V2.5-Pro |
| LMArena Text: Polish | 1439 | 1464 | MiMo-V2.5-Pro |
| LMArena Text: multi-turn | 1440 | 1476 | MiMo-V2.5-Pro |
| LMArena Text: longer queries | 1448 | 1481 | MiMo-V2.5-Pro |
| LMArena Text: writing and language | 1405 | 1451 | MiMo-V2.5-Pro |
| LMArena Text: entertainment and media | 1393 | 1438 | MiMo-V2.5-Pro |
| LMArena Vision: English | 1267 | Not measured | Only Qwen3.8 27B measured |
| LMArena Vision: Chinese | 1326 | Not measured | Only Qwen3.8 27B measured |
| LMArena Vision: reading text | 1281 | Not measured | Only Qwen3.8 27B measured |
| LMArena Vision: humor | 1235 | Not measured | Only Qwen3.8 27B measured |
| LMArena Vision: creative writing | 1237 | Not measured | Only Qwen3.8 27B measured |
| Neither model has a result on the other 15 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Qwen3.8 27B | MiMo-V2.5-Pro | Leader |
|---|---|---|---|
| Neither model has a result on 5 boards here | |||
What each workload costs
Three fixed token mixes at each model's listed price, so per-token rates become a bill.
Chat turn
1K input + 500 output tokens
MiMo-V2.5-Pro costs less
Repository review
50K input + 3K output tokens
Qwen3.8 27B costs less
Agent loop
200K input + 10K output tokens
Qwen3.8 27B costs less
Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.
Specification differences
The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.
| Specification | Qwen3.8 27B | Xiaomi MiMo-V2.5-Pro |
|---|---|---|
| Maker | Qwen | Xiaomi |
| Weights | Open weights | Open weights |
| Released | August 14, 2026 | April 22, 2026 |
| Context window | 1M | 1.05M |
| Input $/Mtok | $0.04 | $0.3045 |
| Output $/Mtok | $1.35 | $0.609 |
| Coding index | 73.6 | 40.0 |
| Independent measurements behind it | 2 | 1 |
| Cost per solved task | $1.26 | $0.03 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Qwen3.8 27B better than MiMo-V2.5-Pro for coding?
Pick Qwen3.8 27B for coding: it scores 73.6 on the Vibe Coding Index against 40.0 for MiMo-V2.5-Pro. Pick MiMo-V2.5-Pro when output tokens cost 2.2x less, or when a solved task costs $0.03 against $1.26.
Which is cheaper, Qwen3.8 27B or MiMo-V2.5-Pro?
Qwen3.8 27B costs $0.04 per million input tokens and $1.35 per million output tokens; MiMo-V2.5-Pro costs $0.3045 and $0.609. Per solved SWE-bench Verified task, Qwen3.8 27B costs $1.26 and MiMo-V2.5-Pro $0.03.
Which has the bigger context window, Qwen3.8 27B or MiMo-V2.5-Pro?
MiMo-V2.5-Pro, with 1.05M tokens against 1M.
How many benchmarks is the Qwen3.8 27B vs MiMo-V2.5-Pro comparison based on?
56 benchmarks have a result for both models as of October 21, 2026. Qwen3.8 27B's coding index rests on 2 independent benchmarks and MiMo-V2.5-Pro's on 1. Where only one model has a result, the page says so instead of naming a winner.