Model comparison
Kimi K2.6 vs Qwen3.8 27B
Kimi K2.6 is the stronger model overall, scoring 47.7 to 46.0 on the Noometry Index. Qwen3.8 27B costs 1.5× less per token, which makes it the better buy when Kimi K2.6's lead doesn't matter for your workload.
Last verified . 27 shared benchmarks.
Alibaba (Qwen)Summary
- They share 27 benchmarks with published results for both. Kimi K2.6 scores higher in 7 categories and Qwen3.8 27B in 3 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K2.6 leads 57.0 to 37.1.
- The biggest single-benchmark swing is NYT Connections (extended): 87.2% for Kimi K2.6 and 54.5% for Qwen3.8 27B.
- Qwen3.8 27B is cheaper at $0.99 / $1.49 per million input/output tokens, against $0.95 / $4 for Kimi K2.6.
Side by side
| Kimi K2.6 | Qwen3.8 27B | |
|---|---|---|
| Provider | Moonshot AI | Alibaba (Qwen) |
| Noometry Index | 47.7 | 46.0 |
| Released | 2026-04-20 | 2026-08-14 |
| Weights | Open | Open |
| Context window | 262K | 262K |
| Max output | 262K | 33K |
| Input $ / M tokens | $0.95 | $0.99 |
| Output $ / M tokens | $4 | $1.49 |
| Results tracked | 51 | 31 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Kimi K2.6: 50.7 (#43), Qwen3.8 27B: 50.5 (#44)
| Benchmark | Kimi K2.6 | Qwen3.8 27B |
|---|---|---|
| LMArena WebDev | 1509 | 1593 |
| SciCode | 53.5% | 46.6% |
| LMArena Coding | 1488 | 1482 |
| SWE-bench Verified | 76.7% | — |
| WeirdML | 55.9% | — |
| ALE-Bench | 1,093 | — |
Agentic & Tool Use Qwen3.8 27B leads
Kimi K2.6: 21.9 (#137), Qwen3.8 27B: 32.9 (#57)
| Benchmark | Kimi K2.6 | Qwen3.8 27B |
|---|---|---|
| APEX-Agents | — | 47.5% |
| OSWorld 2.0 | 4.6% | — |
| ExploitBench | 18.4% | — |
| GBAEval | 0.9% | — |
| GDP.pdf | 12% | — |
| Vending-Bench 2 | 6,205 | — |
Reasoning Too close to call
Kimi K2.6: 40.5 (#55), Qwen3.8 27B: 41.0 (#54)
| Benchmark | Kimi K2.6 | Qwen3.8 27B |
|---|---|---|
| NYT Connections (extended) | 87.2% | 54.5% |
| CritPt | 8% | 5.4% |
| LMArena Hard Prompts | 1470 | 1460 |
| DTBench | 90.9% | 88% |
| LMCA | 37.3% | 41.4% |
| Epoch Capabilities Index | 151.05 | 149.38 |
| ARC-AGI-2 | — | 42.4% |
| ARC-AGI-1 | — | 87.5% |
| Chess Puzzles | 26% | — |
| EBR-Bench | 2.4% | — |
| Mystery Game Puzzles | 18% | — |
| Surface Evolver Bench | — | 45% |
Math Kimi K2.6 leads
Kimi K2.6: 57.0 (#41), Qwen3.8 27B: 37.1 (#161)
| Benchmark | Kimi K2.6 | Qwen3.8 27B |
|---|---|---|
| ProofBench | 16% | 16% |
| LMArena Math | 1475 | 1456 |
| FrontierMath (Tiers 1-3) | 57.2% | — |
| FrontierMath Tier 4 | 25.6% | — |
| MathArena Final-Answer Competitions | 72.9% | — |
| OTIS Mock AIME 2024-2025 | 96.1% | — |
| FrontierMath (Feb 2025 set) | 39% | — |
| FrontierMath Tier 4 (v1) | 14.6% | — |
Knowledge Kimi K2.6 leads
Kimi K2.6: 54.0 (#54), Qwen3.8 27B: 41.6 (#109)
| Benchmark | Kimi K2.6 | Qwen3.8 27B |
|---|---|---|
| LMArena Expert | 1491 | 1482 |
| GPQA Diamond | 90.8% | — |
| SimpleQA Verified | 34.9% | — |
| Vectara Hallucination Rate | 10.8% | — |
Multimodal Qwen3.8 27B leads
Kimi K2.6: 31.6 (#103), Qwen3.8 27B: 41.3 (#37)
| Benchmark | Kimi K2.6 | Qwen3.8 27B |
|---|---|---|
| LMArena Vision | 1283 | 1271 |
| Blueprint-Bench 2 | 3.9% | — |
| Furniture Assembly | 21.7% | — |
| LMArena Document | 1451 | — |
Multilingual Kimi K2.6 leads
Kimi K2.6: 54.9 (#37), Qwen3.8 27B: 53.7 (#60)
| Benchmark | Kimi K2.6 | Qwen3.8 27B |
|---|---|---|
| LMArena Non-English | 1446 | 1430 |
| LMArena Chinese | 1521 | 1504 |
| LMArena French | 1471 | 1465 |
| LMArena German | 1450 | 1438 |
| LMArena Japanese | 1443 | 1384 |
| LMArena Korean | 1427 | 1393 |
| LMArena Russian | 1446 | 1415 |
| LMArena Spanish | 1464 | 1448 |
Instruction Following Too close to call
Kimi K2.6: 76.3 (#43), Qwen3.8 27B: 75.8 (#53)
| Benchmark | Kimi K2.6 | Qwen3.8 27B |
|---|---|---|
| LMArena Instruction Following | 1451 | 1439 |
Long Context Too close to call
Kimi K2.6: 44.9 (#52), Qwen3.8 27B: 44.3 (#70)
| Benchmark | Kimi K2.6 | Qwen3.8 27B |
|---|---|---|
| LMArena Longer Query | 1468 | 1450 |
Writing & Preference Kimi K2.6 leads
Kimi K2.6: 68.5 (#26), Qwen3.8 27B: 65.8 (#43)
| Benchmark | Kimi K2.6 | Qwen3.8 27B |
|---|---|---|
| LMArena Text | 1455 | 1441 |
| LMArena Creative Writing | 1434 | 1384 |
| EQ-Bench Creative Writing | 1725 | 1671 |
| LMArena Multi-Turn | 1453 | 1441 |
| EQ-Bench 4 | 1202 | — |
Frequently asked questions
Is Kimi K2.6 better than Qwen3.8 27B?
Kimi K2.6 is the stronger model overall, scoring 47.7 to 46.0 on the Noometry Index. Qwen3.8 27B costs 1.5× less per token, which makes it the better buy when Kimi K2.6's lead doesn't matter for your workload.
Which is cheaper, Kimi K2.6 or Qwen3.8 27B?
Qwen3.8 27B is cheaper. It lists at $0.99 per million input tokens and $1.49 per million output tokens; Kimi K2.6 lists at $0.95 and $4.
Is Kimi K2.6 or Qwen3.8 27B better for coding?
They score almost the same on coding (50.7 vs 50.5); test both on your own repository before choosing.
Which has the bigger context window?
Both accept 262K tokens.
How many benchmarks do Kimi K2.6 and Qwen3.8 27B share?
27 benchmarks have published results for both models. Kimi K2.6 has 51 scored results on Noometry and Qwen3.8 27B has 31.