Mistral Medium 3.5 vs Qwen3.8 27B
Updated October 21, 2026. 65 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 9.2 for Mistral Medium 3.5.
Pick Mistral Medium 3.5 when what matters is MedCode (Vals AI) or LMArena Text: German, where it leads.
- Shared results
- 65
- Mistral Medium 3.5 only
- 15
- Qwen3.8 27B only
- 20
- 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 9.2.
1 and 2 benchmarks behind the scores
Intelligence
AA Intelligence Index
Qwen3.8 27B
Qwen3.8 27B leads, 33.7 against 14.2.
Same board, both measured
Agentic work
AA Agentic Index
Qwen3.8 27B
Qwen3.8 27B leads, 45.8 against 7.0.
Same board, both measured
Reasoning
GPQA Diamond
Qwen3.8 27B
Qwen3.8 27B leads, 88.9% against 34.9%.
Same board, both measured
Knowledge
MMLU Pro
Qwen3.8 27B
Qwen3.8 27B leads, 84.3% against 75.3%.
Same board, both measured
Web design
Design Arena: websites
No call
Only Mistral Medium 3.5 has a Design Arena: websites result, 1142.
Directional only
Long documents
Listed context window
Qwen3.8 27B
Qwen3.8 27B takes 1M tokens against 262K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Qwen3.8 27B
Mistral Medium 3.5 $0.0053, Qwen3.8 27B $0.0007 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
Qwen3.8 27B
Mistral Medium 3.5 $6.02, Qwen3.8 27B $1.26.
Published attempt cost and pass rate
At a glance
- Coding index
- Mistral Medium 3.5 9.2 from 1 benchmark, Qwen3.8 27B 73.6 from 2 benchmarks.
- Widest category
- Agentic work, 31.4 points apart on average over 5 shared benchmarks.
- Widest single gap
- Vibe Code Bench (Vals AI): Mistral Medium 3.5 2.9% against Qwen3.8 27B 64.8%, 62.0 points apart on a 0 to 100 scale.
- Output price
- Mistral Medium 3.5 $7.50 against Qwen3.8 27B $1.35 per million output tokens, 5.6x apart.
- Context window
- Mistral Medium 3.5 takes 262K tokens, Qwen3.8 27B 1M.
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 | Mistral Medium 3.5 | Bars | Qwen3.8 27B | Ties |
|---|---|---|---|---|
| Coding | 0 | 11 | ||
| Intelligence | 0 | 2 | ||
| Agentic work | 0 | 5 | ||
| Reasoning | 0 | 8 | ||
| Math | 0 | 3 | ||
| Knowledge | 1 | 7 | ||
| General | 1 | 27 | ||
| All shared boards | 2 | 63 |
CodingLike-for-likeFour of these boards feed the Vibe Coding Index011
| Benchmark | Mistral Medium 3.5 | Qwen3.8 27B | Leader |
|---|---|---|---|
| LiveBench coding | Not measured | 86.8% | Only Qwen3.8 27B measured |
| LMArena WebDev | 1264 | 1593 | Qwen3.8 27B |
| AA Coding Index | 46.9 | 68.1 | Qwen3.8 27B |
| LiveCodeBench (Vals AI) | Not measured | 84.0% | Only Qwen3.8 27B measured |
| SWE-bench Verified (Vals AI) | 66.4% | 86.0% | Qwen3.8 27B |
| Vibe Code Bench (Vals AI) | 2.9% | 64.8% | Qwen3.8 27B |
| Code migration (Vals AI) | 5.1% | 14.2% | Qwen3.8 27B |
| ProgramBench (Vals AI) | Not measured | 0.0% | Only Qwen3.8 27B measured |
| IOI (Vals AI) | Not measured | 39.1% | Only Qwen3.8 27B measured |
| Terminal-Bench 2.1 (Vals AI) | 39.0% | 58.4% | Qwen3.8 27B |
| SkillsBench (Vals AI) | Not measured | 38.1% | Only Qwen3.8 27B measured |
| CyberBench (Vals AI) | Not measured | 67.8% | Only Qwen3.8 27B measured |
| SciCode | 40.2% | 46.6% | Qwen3.8 27B |
| Terminal-Bench Hard (Artificial Analysis run) | 33.3% | Not measured | Only Mistral Medium 3.5 measured |
| ALE-Bench | 764.0 | Not measured | Only Mistral Medium 3.5 measured |
| LMArena Text: coding | 1460 | 1483 | Qwen3.8 27B |
| LMArena Text: software and IT | 1458 | 1478 | Qwen3.8 27B |
| LMArena WebDev: React | 1254 | 1602 | Qwen3.8 27B |
| LMArena WebDev: HTML | 1308 | 1570 | Qwen3.8 27B |
| LMArena WebDev: from an image | Not measured | 1577 | Only Qwen3.8 27B measured |
| Neither model has a result on the other 18 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
| Benchmark | Mistral Medium 3.5 | Qwen3.8 27B | Leader |
|---|---|---|---|
| AA Intelligence Index | 14.2 | 33.7 | Qwen3.8 27B |
| Epoch Capabilities Index | 141.4 | 151.7 | Qwen3.8 27B |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index05
| Benchmark | Mistral Medium 3.5 | Qwen3.8 27B | Leader |
|---|---|---|---|
| AA Agentic Index | 7.0 | 45.8 | Qwen3.8 27B |
| LiveBench research | Not measured | 59.4% | Only Qwen3.8 27B measured |
| LiveBench browser | Not measured | 32.0% | Only Qwen3.8 27B measured |
| LiveBench long agentic | Not measured | 27.9% | Only Qwen3.8 27B measured |
| Finance Agent (Vals AI) | 32.1% | 48.5% | Qwen3.8 27B |
| Tax Agent Bench (Vals AI) | 4.5% | 30.9% | Qwen3.8 27B |
| Legal Research Bench (Vals AI) | 9.1% | 36.1% | Qwen3.8 27B |
| Excel Modeling Benchmark (Vals AI) | 11.2% | 59.7% | Qwen3.8 27B |
| Terminal-Bench-Science (Vals AI) | Not measured | 0.0% | Only Qwen3.8 27B measured |
| τ²-Bench Telecom | 94.2% | Not measured | Only Mistral Medium 3.5 measured |
| APEX-Agents | Not measured | 47.5% | 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 index08
| Benchmark | Mistral Medium 3.5 | Qwen3.8 27B | Leader |
|---|---|---|---|
| GPQA Diamond | 34.9% | 88.9% | Qwen3.8 27B |
| LiveBench reasoning | Not measured | 80.0% | Only Qwen3.8 27B measured |
| MMMU Pro (Vals AI) | Not measured | 83.9% | Only Qwen3.8 27B measured |
| GPQA Diamond (Artificial Analysis run) | 74.8% | 90.5% | Qwen3.8 27B |
| Humanity's Last Exam (Artificial Analysis run) | 13.8% | 33.9% | Qwen3.8 27B |
| ARC-AGI | Not measured | 87.5% | Only Qwen3.8 27B measured |
| ARC-AGI-2 | Not measured | 42.4% | Only Qwen3.8 27B measured |
| LMArena Text: hard prompts | 1436 | 1459 | Qwen3.8 27B |
| LMArena Text: expert | 1432 | 1479 | Qwen3.8 27B |
| LMArena Text: hard prompts in English | 1445 | 1461 | Qwen3.8 27B |
| LMArena Vision: diagrams | 1254 | 1285 | Qwen3.8 27B |
| LMArena Vision: homework | 1230 | 1295 | Qwen3.8 27B |
| Neither model has a result on the other 9 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index03
| Benchmark | Mistral Medium 3.5 | Qwen3.8 27B | Leader |
|---|---|---|---|
| ProofBench | 9.0% | 16.0% | Qwen3.8 27B |
| LMArena Text: math | 1431 | 1455 | Qwen3.8 27B |
| LMArena Text: mathematical work | 1440 | 1458 | Qwen3.8 27B |
| Neither model has a result on the other 10 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index17
| Benchmark | Mistral Medium 3.5 | Qwen3.8 27B | Leader |
|---|---|---|---|
| MMLU Pro | 75.3% | 84.3% | Qwen3.8 27B |
| LegalBench (Vals AI) | Not measured | 82.4% | Only Qwen3.8 27B measured |
| CaseLaw (Vals AI) | 44.2% | Not measured | Only Mistral Medium 3.5 measured |
| CorpFin (Vals AI) | 58.8% | Not measured | Only Mistral Medium 3.5 measured |
| TaxEval (Vals AI) | 68.0% | 70.8% | Qwen3.8 27B |
| MortgageTax (Vals AI) | 28.9% | 64.9% | Qwen3.8 27B |
| MedCode (Vals AI) | 33.8% | 28.7% | Mistral Medium 3.5 |
| LMArena Text: science | 1428 | 1467 | Qwen3.8 27B |
| LMArena Text: medicine and healthcare | 1409 | 1437 | Qwen3.8 27B |
| LMArena Text: legal and government | 1437 | 1447 | Qwen3.8 27B |
| LMArena Text: business and finance | 1426 | 1438 | Qwen3.8 27B |
| Neither model has a result on the other 6 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index9 boards, none shared
| Benchmark | Mistral Medium 3.5 | Qwen3.8 27B | Leader |
|---|---|---|---|
| Design Arena: websites | 1142 | Not measured | Only Mistral Medium 3.5 measured |
| Design Arena: UI components | 1079 | Not measured | Only Mistral Medium 3.5 measured |
| Design Arena: game dev | 1079 | Not measured | Only Mistral Medium 3.5 measured |
| Design Arena: data viz | 1120 | Not measured | Only Mistral Medium 3.5 measured |
| Design Arena: 3D | 1061 | Not measured | Only Mistral Medium 3.5 measured |
| Design Arena: all code categories | 1122 | Not measured | Only Mistral Medium 3.5 measured |
| Design Arena: SVG | 949 | Not measured | Only Mistral Medium 3.5 measured |
| Design Arena: ASCII art | 1062 | Not measured | Only Mistral Medium 3.5 measured |
| LiveBench design | Not measured | 75.9% | Only Qwen3.8 27B measured |
GeneralLike-for-likePublic boards, shown as published, kept out of the index127
| Benchmark | Mistral Medium 3.5 | Qwen3.8 27B | Leader |
|---|---|---|---|
| LiveBench data analysis | Not measured | 76.6% | Only Qwen3.8 27B measured |
| SAGE (Vals AI) | 37.6% | 52.4% | Qwen3.8 27B |
| MedScribe (Vals AI) | 67.7% | 83.8% | Qwen3.8 27B |
| Vals Multimodal Index | 34.8% | Not measured | Only Mistral Medium 3.5 measured |
| IFBench | 68.8% | Not measured | Only Mistral Medium 3.5 measured |
| Long context reasoning (AA-LCR) | 69.3% | 82.0% | Qwen3.8 27B |
| DTBench | 75.5% | 88.0% | Qwen3.8 27B |
| LMCA | 26.1% | 41.4% | Qwen3.8 27B |
| Surface Evolver Bench | 26.9% | 45.0% | Qwen3.8 27B |
| LMArena Text | 1421 | 1440 | Qwen3.8 27B |
| LMArena Text: instruction following | 1415 | 1436 | Qwen3.8 27B |
| LMArena Text: creative writing | 1374 | 1383 | Qwen3.8 27B |
| LMArena Vision | 1222 | 1269 | Qwen3.8 27B |
| LMArena Text: English | 1433 | 1448 | Qwen3.8 27B |
| LMArena Text: non-English | 1404 | 1428 | Qwen3.8 27B |
| LMArena Text: Chinese | 1444 | 1501 | Qwen3.8 27B |
| LMArena Text: German | 1453 | 1443 | Mistral Medium 3.5 |
| LMArena Text: Japanese | Not measured | 1387 | Only Qwen3.8 27B measured |
| LMArena Text: Korean | 1384 | 1384 | Qwen3.8 27B |
| LMArena Text: French | 1448 | 1465 | Qwen3.8 27B |
| LMArena Text: Spanish | 1411 | 1449 | Qwen3.8 27B |
| LMArena Text: Russian | 1395 | 1415 | Qwen3.8 27B |
| LMArena Text: Polish | 1430 | 1439 | Qwen3.8 27B |
| LMArena Text: multi-turn | 1423 | 1440 | Qwen3.8 27B |
| LMArena Text: longer queries | 1415 | 1448 | Qwen3.8 27B |
| LMArena Text: writing and language | 1384 | 1405 | Qwen3.8 27B |
| LMArena Text: entertainment and media | 1382 | 1393 | Qwen3.8 27B |
| LMArena Vision: English | 1240 | 1267 | Qwen3.8 27B |
| LMArena Vision: Chinese | 1244 | 1326 | Qwen3.8 27B |
| LMArena Vision: reading text | 1233 | 1281 | Qwen3.8 27B |
| LMArena Vision: humor | 1216 | 1235 | Qwen3.8 27B |
| LMArena Vision: creative writing | 1198 | 1237 | Qwen3.8 27B |
| Neither model has a result on the other 14 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Mistral Medium 3.5 | Qwen3.8 27B | 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
Qwen3.8 27B 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 | Mistral Medium 3.5 | Qwen3.8 27B |
|---|---|---|
| Maker | Mistral AI | Qwen |
| Weights | No public weights listed | Open weights |
| Released | April 30, 2026 | August 14, 2026 |
| Context window | 262K | 1M |
| Input $/Mtok | $1.50 | $0.04 |
| Output $/Mtok | $7.50 | $1.35 |
| Coding index | 9.2 | 73.6 |
| Independent measurements behind it | 1 | 2 |
| Cost per solved task | $6.02 | $1.26 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Mistral Medium 3.5 better than Qwen3.8 27B for coding?
Pick Qwen3.8 27B for coding: it scores 73.6 on the Vibe Coding Index against 9.2 for Mistral Medium 3.5. Pick Mistral Medium 3.5 when what matters is MedCode (Vals AI) or LMArena Text: German, where it leads.
Which is cheaper, Mistral Medium 3.5 or Qwen3.8 27B?
Mistral Medium 3.5 costs $1.50 per million input tokens and $7.50 per million output tokens; Qwen3.8 27B costs $0.04 and $1.35. Per solved SWE-bench Verified task, Mistral Medium 3.5 costs $6.02 and Qwen3.8 27B $1.26.
Which has the bigger context window, Mistral Medium 3.5 or Qwen3.8 27B?
Qwen3.8 27B, with 1M tokens against 262K.
How many benchmarks is the Mistral Medium 3.5 vs Qwen3.8 27B comparison based on?
65 benchmarks have a result for both models as of October 21, 2026. Mistral Medium 3.5's coding index rests on 1 independent benchmark and Qwen3.8 27B's on 2. Where only one model has a result, the page says so instead of naming a winner.