Qwen3.8-27BvsQwen3.6-27B
Across 4 shared benchmarks, Qwen3.8-27B leads overall: Qwen3.8-27B wins 4, Qwen3.6-27B wins 0, with 0 ties and an average score difference of +5.70.
Qwen3.8-27B
阿里巴巴 · 2026-08-14 · Reasoning model
Qwen3.6-27B
阿里巴巴 · 2026-04-22 · Reasoning model
Qwen3.8-27B4 wins(100%)(0%)0 winsQwen3.6-27B
Benchmark scores
Grouped by capability, sorted by largest gap within each. 4 shared benchmarks.
Coding and Software Engineer
Qwen3.8-27B 2/2| Benchmark | Qwen3.8-27B | Qwen3.6-27B | Diff |
|---|---|---|---|
| SWE-Bench Pro - Public | 61.7010 / 57Thinking (With Tools) | 53.5036 / 57Thinking (With Tools) | +8.20 |
| LiveCodeBench | 90.306 / 124Thinking (No Tools) | 83.9020 / 124Thinking (No Tools) | +6.40 |
General Evaluation
Qwen3.8-27B 1/1| Benchmark | Qwen3.8-27B | Qwen3.6-27B | Diff |
|---|---|---|---|
| GPQA Diamond | 89.2047 / 226Thinking (No Tools) | 87.8064 / 226Thinking (No Tools) | +1.40 |
General Knowledge
Qwen3.8-27B 1/1| Benchmark | Qwen3.8-27B | Qwen3.6-27B | Diff |
|---|---|---|---|
| HLE | 30.8087 / 179Thinking (No Tools) | 24113 / 179Thinking (No Tools) | +6.80 |
Specs
| Field | Qwen3.8-27B | Qwen3.6-27B |
|---|---|---|
| Publisher | 阿里巴巴 | 阿里巴巴 |
| Release date | 2026-08-14 | 2026-04-22 |
| Model type | Reasoning model | Reasoning model |
| Architecture | Dense | Dense |
| Parameters | 27B | 27B |
| Context length | 256K | 128K |
| Max output | 128K | 16K |
Summary
- Qwen3.8-27Bleads in:Coding and Software Engineer (2/2), General Evaluation (1/1), General Knowledge (1/1)
On average across the 4 shared benchmarks, Qwen3.8-27B scores 5.70 higher.
Largest single-benchmark gap: SWE-Bench Pro - Public — Qwen3.8-27B 61.70 vs Qwen3.6-27B 53.50 (+8.20).
Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.