Kimi K3 vs Qwen3.8 Max (0902)
Updated October 21, 2026. 101 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.
Decision reading
Kimi K3 and Qwen3.8 Max (0902) are level for coding, 61.8 against 63.5 on the Vibe Coding Index, so price and context decide.
Pick Kimi K3 when a solved task costs $0.81 against $1.32, or when the context window is larger, 1.05M against 1M tokens. Pick Qwen3.8 Max (0902) when output tokens cost 2.0x less, or when what matters is LiveBench coding or LMArena WebDev, where it leads.
- Shared results
- 101
- Kimi K3 only
- 16
- Qwen3.8 Max (0902) only
- 12
- Like-for-like categories
- 8 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
No call
Level: 61.8 against 63.5, inside 2 points.
2 benchmarks behind each score
Intelligence
AA Intelligence Index
Qwen3.8 Max (0902)
Qwen3.8 Max (0902) leads, 45.4 against 43.6.
Same board, both measured
Agentic work
AA Agentic Index
Qwen3.8 Max (0902)
Qwen3.8 Max (0902) leads, 56.0 against 50.0.
Same board, both measured
Reasoning
GPQA Diamond
Qwen3.8 Max (0902)
Qwen3.8 Max (0902) leads, 93.7% against 92.9%.
Same board, both measured
Knowledge
MMLU Pro
Qwen3.8 Max (0902)
Qwen3.8 Max (0902) leads, 88.6% against 88.0%.
Same board, both measured
Web design
Design Arena: websites
Kimi K3
Kimi K3 leads, 1336 against 1275.
Same board, both measured
General
Vals Index
No call
Only Kimi K3 has a Vals Index result, 53.0%.
Directional only
Long documents
Listed context window
Kimi K3
Kimi K3 takes 1.05M tokens against 1M.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Qwen3.8 Max (0902)
Kimi K3 $0.0062, Qwen3.8 Max (0902) $0.0050 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
Kimi K3
Kimi K3 $0.81, Qwen3.8 Max (0902) $1.32.
Published attempt cost and pass rate
At a glance
- Coding index
- Kimi K3 61.8 from 2 benchmarks, Qwen3.8 Max (0902) 63.5 from 2 benchmarks.
- Widest category
- Coding, 7.9 points apart on average over 21 shared benchmarks.
- Widest single gap
- CyberBench (Vals AI): Kimi K3 75.2% against Qwen3.8 Max (0902) 28.6%, 46.7 points apart on a 0 to 100 scale.
- Output price
- Kimi K3 $11.936 against Qwen3.8 Max (0902) $6.00 per million output tokens, 2.0x apart.
- Context window
- Kimi K3 takes 1.05M tokens, Qwen3.8 Max (0902) 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 | Kimi K3 | Bars | Qwen3.8 Max (0902) | Ties |
|---|---|---|---|---|
| Coding | 11 | 9 | 1 tie | |
| Intelligence | 1 | 1 | ||
| Agentic work | 10 | 7 | ||
| Reasoning | 6 | 5 | ||
| Math | 2 | 4 | ||
| Knowledge | 7 | 4 | ||
| Web design | 6 | 1 | ||
| General | 13 | 13 | ||
| All shared boards | 56 | 44 | 1 tie |
CodingLike-for-likeFour of these boards feed the Vibe Coding Index119
| Benchmark | Kimi K3 | Qwen3.8 Max (0902) | Leader |
|---|---|---|---|
| LiveBench coding | 59.3% | 60.3% | Qwen3.8 Max (0902) |
| LMArena WebDev | 1655 | 1674 | Qwen3.8 Max (0902) |
| AA Coding Index | 76.2 | 76.2 | Tie |
| LiveCodeBench (Vals AI) | 87.2% | 87.8% | Qwen3.8 Max (0902) |
| SWE-bench Verified (Vals AI) | 93.4% | 85.6% | Kimi K3 |
| Terminal-Bench 4 (Vals AI) | 17.2% | 34.3% | Qwen3.8 Max (0902) |
| Vibe Code Bench (Vals AI) | 85.0% | 64.7% | Kimi K3 |
| Code migration (Vals AI) | 16.1% | 24.0% | Qwen3.8 Max (0902) |
| ProgramBench (Vals AI) | 2.0% | 0.0% | Kimi K3 |
| IOI (Vals AI) | 48.9% | 68.9% | Qwen3.8 Max (0902) |
| Terminal-Bench 2.1 (Vals AI) | 80.9% | 67.4% | Kimi K3 |
| SWE-bench Multimodal (Vals AI) | 44.2% | 42.5% | Kimi K3 |
| Vibe Code Bench 1-100 (Vals AI) | 18.2% | 12.8% | Kimi K3 |
| SkillsBench (Vals AI) | Not measured | 42.0% | Only Qwen3.8 Max (0902) measured |
| CyberBench (Vals AI) | 75.2% | 28.6% | Kimi K3 |
| SciCode | 59.5% | 53.2% | Kimi K3 |
| FrontierCode | 44.2% | Not measured | Only Kimi K3 measured |
| FrontierSWE | 25.9% | 17.8% | Kimi K3 |
| WeirdML | 82.6% | Not measured | Only Kimi K3 measured |
| ALE-Bench | 1524.5 | Not measured | Only Kimi K3 measured |
| LMArena Text: coding | 1508 | 1502 | Kimi K3 |
| LMArena Text: software and IT | 1499 | 1498 | Kimi K3 |
| LMArena WebDev: React | 1670 | 1692 | Qwen3.8 Max (0902) |
| LMArena WebDev: HTML | 1644 | 1649 | Qwen3.8 Max (0902) |
| LMArena WebDev: from an image | 1579 | 1637 | Qwen3.8 Max (0902) |
| Neither model has a result on the other 13 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index11
| Benchmark | Kimi K3 | Qwen3.8 Max (0902) | Leader |
|---|---|---|---|
| AA Intelligence Index | 43.6 | 45.4 | Qwen3.8 Max (0902) |
| Epoch Capabilities Index | 157.9 | 157.4 | Kimi K3 |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index107
| Benchmark | Kimi K3 | Qwen3.8 Max (0902) | Leader |
|---|---|---|---|
| AA Agentic Index | 50.0 | 56.0 | Qwen3.8 Max (0902) |
| Design Arena: agents building web apps | 1296 | 1301 | Qwen3.8 Max (0902) |
| Design Arena: agents building full-stack apps | 1315 | 1305 | Kimi K3 |
| Design Arena: agents building mobile apps | 1267 | 1223 | Kimi K3 |
| Design Arena: agents building Android apps | 1256 | 1207 | Kimi K3 |
| Design Arena: agents building Godot games | 1223 | 1234 | Qwen3.8 Max (0902) |
| Design Arena: agents building games | 1252 | Not measured | Only Kimi K3 measured |
| Design Arena: HTML slides | 1256 | Not measured | Only Kimi K3 measured |
| Design Arena: PowerPoint slides | 1273 | 1257 | Kimi K3 |
| Design Arena: spreadsheets | 1058 | 868 | Kimi K3 |
| LiveBench research | 79.0% | 69.9% | Kimi K3 |
| LiveBench browser | 16.0% | 32.0% | Qwen3.8 Max (0902) |
| LiveBench long agentic | 42.8% | 42.3% | Kimi K3 |
| Finance Agent (Vals AI) | 53.1% | 50.6% | Kimi K3 |
| Tax Agent Bench (Vals AI) | 33.4% | 32.1% | Kimi K3 |
| Legal Research Bench (Vals AI) | 46.1% | 47.6% | Qwen3.8 Max (0902) |
| Harvey's Legal Agent Benchmark (Vals AI) | 12.9% | Not measured | Only Kimi K3 measured |
| Excel Modeling Benchmark (Vals AI) | 66.7% | 60.1% | Kimi K3 |
| BioMysteryBench (Vals AI) | 72.6% | Not measured | Only Kimi K3 measured |
| Terminal-Bench-Science (Vals AI) | 1.4% | 12.9% | Qwen3.8 Max (0902) |
| Time Horizon Index (Vals AI) | 10.5% | Not measured | Only Kimi K3 measured |
| APEX-Agents | 47.8% | 63.3% | Qwen3.8 Max (0902) |
| PostTrainBench | 32.0% | Not measured | Only Kimi K3 measured |
| Vending-Bench 2 | $5,165 | Not measured | Only Kimi K3 measured |
| LMArena Search | Not measured | 1261 | Only Qwen3.8 Max (0902) measured |
| Neither model has a result on the other 14 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index65
| Benchmark | Kimi K3 | Qwen3.8 Max (0902) | Leader |
|---|---|---|---|
| GPQA Diamond | 92.9% | 93.7% | Qwen3.8 Max (0902) |
| LiveBench reasoning | 90.7% | 88.2% | Kimi K3 |
| MysteryMechanism (Vals AI) | Not measured | 23.9% | Only Qwen3.8 Max (0902) measured |
| MMMU Pro (Vals AI) | 88.2% | 88.0% | Kimi K3 |
| GPQA Diamond (Artificial Analysis run) | 93.5% | 92.8% | Kimi K3 |
| Humanity's Last Exam (Artificial Analysis run) | 46.9% | 43.1% | Kimi K3 |
| GPQA Diamond (Epoch AI run) | 93.1% | 92.7% | Kimi K3 |
| ARC-AGI | 94.5% | Not measured | Only Kimi K3 measured |
| ARC-AGI-2 | 60.4% | Not measured | Only Kimi K3 measured |
| SimpleBench | 60.7% | Not measured | Only Kimi K3 measured |
| Chess puzzles | 37.0% | 46.0% | Qwen3.8 Max (0902) |
| Mystery game puzzles | 26.0% | 38.0% | Qwen3.8 Max (0902) |
| LMArena Text: hard prompts | 1495 | 1496 | Qwen3.8 Max (0902) |
| LMArena Text: expert | 1520 | 1507 | Kimi K3 |
| LMArena Text: hard prompts in English | 1496 | 1503 | Qwen3.8 Max (0902) |
| LMArena Vision: diagrams | Not measured | 1323 | Only Qwen3.8 Max (0902) measured |
| LMArena Vision: homework | Not measured | 1336 | Only Qwen3.8 Max (0902) measured |
| Neither model has a result on the other 4 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index24
| Benchmark | Kimi K3 | Qwen3.8 Max (0902) | Leader |
|---|---|---|---|
| OTIS Mock AIME | 97.2% | 100.0% | Qwen3.8 Max (0902) |
| FrontierMath, tiers 1 to 3 | 72.2% | 74.7% | Qwen3.8 Max (0902) |
| FrontierMath, tier 4 | 39.0% | 46.3% | Qwen3.8 Max (0902) |
| ProofBench | 87.0% | 58.0% | Kimi K3 |
| LMArena Text: math | 1494 | 1501 | Qwen3.8 Max (0902) |
| LMArena Text: mathematical work | 1505 | 1491 | Kimi K3 |
| Neither model has a result on the other 7 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index74
| Benchmark | Kimi K3 | Qwen3.8 Max (0902) | Leader |
|---|---|---|---|
| MMLU Pro | 88.0% | 88.6% | Qwen3.8 Max (0902) |
| LegalBench (Vals AI) | 86.2% | 83.6% | Kimi K3 |
| CorpFin (Vals AI) | 71.6% | 65.8% | Kimi K3 |
| TaxEval (Vals AI) | 75.7% | 75.5% | Kimi K3 |
| MortgageTax (Vals AI) | 66.3% | 64.0% | Kimi K3 |
| MedCode (Vals AI) | 49.4% | 40.7% | Kimi K3 |
| SimpleQA Verified | 50.6% | 47.3% | Kimi K3 |
| LMArena Text: science | 1508 | 1512 | Qwen3.8 Max (0902) |
| LMArena Text: medicine and healthcare | 1489 | 1503 | Qwen3.8 Max (0902) |
| LMArena Text: legal and government | 1503 | 1492 | Kimi K3 |
| LMArena Text: business and finance | 1470 | 1480 | Qwen3.8 Max (0902) |
| Neither model has a result on the other 6 boards here | |||
Web designLike-for-likePublic boards, shown as published, kept out of the index61
| Benchmark | Kimi K3 | Qwen3.8 Max (0902) | Leader |
|---|---|---|---|
| Design Arena: websites | 1336 | 1275 | Kimi K3 |
| Design Arena: UI components | 1345 | 1340 | Kimi K3 |
| Design Arena: game dev | 1361 | 1313 | Kimi K3 |
| Design Arena: data viz | 1350 | 1281 | Kimi K3 |
| Design Arena: 3D | 1388 | 1337 | Kimi K3 |
| Design Arena: all code categories | 1361 | 1293 | Kimi K3 |
| Design Arena: SVG | 1301 | Not measured | Only Kimi K3 measured |
| Design Arena: ASCII art | Not measured | 1201 | Only Qwen3.8 Max (0902) measured |
| LiveBench design | 77.2% | 78.6% | Qwen3.8 Max (0902) |
GeneralLike-for-likePublic boards, shown as published, kept out of the index1313
| Benchmark | Kimi K3 | Qwen3.8 Max (0902) | Leader |
|---|---|---|---|
| Vals Index | 53.0% | Not measured | Only Kimi K3 measured |
| LiveBench data analysis | 78.7% | 78.4% | Kimi K3 |
| Public Benefits Bench (Vals AI) | 68.2% | 67.1% | Kimi K3 |
| SAGE (Vals AI) | 52.8% | 51.2% | Kimi K3 |
| MedScribe (Vals AI) | 88.0% | 85.0% | Kimi K3 |
| Vals Multimodal Index | 73.4% | 65.4% | Kimi K3 |
| Long context reasoning (AA-LCR) | 88.7% | 80.3% | Kimi K3 |
| DTBench | 91.2% | 92.0% | Qwen3.8 Max (0902) |
| LMCA | 52.7% | 46.2% | Kimi K3 |
| Surface Evolver Bench | 95.0% | Not measured | Only Kimi K3 measured |
| Furniture assembly | 34.2% | 20.0% | Kimi K3 |
| LMArena Text | 1476 | 1483 | Qwen3.8 Max (0902) |
| LMArena Text: instruction following | 1485 | 1481 | Kimi K3 |
| LMArena Text: creative writing | 1457 | 1477 | Qwen3.8 Max (0902) |
| LMArena Vision | Not measured | 1313 | Only Qwen3.8 Max (0902) measured |
| LMArena Text: English | 1481 | 1491 | Qwen3.8 Max (0902) |
| LMArena Text: non-English | 1466 | 1473 | Qwen3.8 Max (0902) |
| LMArena Text: Chinese | 1529 | 1539 | Qwen3.8 Max (0902) |
| LMArena Text: German | 1490 | 1478 | Kimi K3 |
| LMArena Text: Japanese | 1494 | 1473 | Kimi K3 |
| LMArena Text: Korean | 1460 | 1460 | Qwen3.8 Max (0902) |
| LMArena Text: French | 1492 | 1498 | Qwen3.8 Max (0902) |
| LMArena Text: Spanish | 1472 | 1489 | Qwen3.8 Max (0902) |
| LMArena Text: Russian | 1485 | 1482 | Kimi K3 |
| LMArena Text: Polish | 1458 | 1471 | Qwen3.8 Max (0902) |
| LMArena Text: multi-turn | 1489 | 1489 | Qwen3.8 Max (0902) |
| LMArena Text: longer queries | 1494 | 1488 | Kimi K3 |
| LMArena Text: writing and language | 1471 | 1477 | Qwen3.8 Max (0902) |
| LMArena Text: entertainment and media | 1453 | 1464 | Qwen3.8 Max (0902) |
| LMArena Vision: English | Not measured | 1311 | Only Qwen3.8 Max (0902) measured |
| LMArena Vision: Chinese | Not measured | 1358 | Only Qwen3.8 Max (0902) measured |
| LMArena Vision: reading text | Not measured | 1322 | Only Qwen3.8 Max (0902) measured |
| LMArena Vision: humor | Not measured | 1304 | Only Qwen3.8 Max (0902) measured |
| LMArena Vision: creative writing | Not measured | 1312 | Only Qwen3.8 Max (0902) measured |
| Neither model has a result on the other 12 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Kimi K3 | Qwen3.8 Max (0902) | 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 Max (0902) costs less
Repository review
50K input + 3K output tokens
Kimi K3 costs less
Agent loop
200K input + 10K output tokens
Kimi K3 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 | MoonshotAI Kimi K3 | Qwen3.8 Max (0902) |
|---|---|---|
| Maker | Moonshot AI | Qwen |
| Weights | Open weights | No public weights listed |
| Released | July 16, 2026 | September 3, 2026 |
| Context window | 1.05M | 1M |
| Input $/Mtok | $0.239 | $2.00 |
| Output $/Mtok | $11.936 | $6.00 |
| Coding index | 61.8 | 63.5 |
| Independent measurements behind it | 2 | 2 |
| Cost per solved task | $0.81 | $1.32 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Kimi K3 better than Qwen3.8 Max (0902) for coding?
Kimi K3 and Qwen3.8 Max (0902) are level for coding, 61.8 against 63.5 on the Vibe Coding Index, so price and context decide. Pick Kimi K3 when a solved task costs $0.81 against $1.32, or when the context window is larger, 1.05M against 1M tokens. Pick Qwen3.8 Max (0902) when output tokens cost 2.0x less, or when what matters is LiveBench coding or LMArena WebDev, where it leads.
Which is cheaper, Kimi K3 or Qwen3.8 Max (0902)?
Kimi K3 costs $0.239 per million input tokens and $11.936 per million output tokens; Qwen3.8 Max (0902) costs $2.00 and $6.00. Per solved SWE-bench Verified task, Kimi K3 costs $0.81 and Qwen3.8 Max (0902) $1.32.
Which has the bigger context window, Kimi K3 or Qwen3.8 Max (0902)?
Kimi K3, with 1.05M tokens against 1M.
How many benchmarks is the Kimi K3 vs Qwen3.8 Max (0902) comparison based on?
101 benchmarks have a result for both models as of October 21, 2026. Kimi K3's coding index rests on 2 independent benchmarks and Qwen3.8 Max (0902)'s on 2. Where only one model has a result, the page says so instead of naming a winner.