Qwen3.8-MaxvsQwen3.6-Max-Preview
Across 3 shared benchmarks, Qwen3.8-Max leads overall: Qwen3.8-Max wins 3, Qwen3.6-Max-Preview wins 0, with 0 ties and an average score difference of +6.20.
Qwen3.8-Max
阿里巴巴 · 2026-08-03 · Reasoning model
Qwen3.6-Max-Preview
阿里巴巴 · 2026-04-18 · Chat model
Qwen3.8-Max3 wins(100%)(0%)0 winsQwen3.6-Max-Preview
Benchmark scores
Grouped by capability, sorted by largest gap within each. 3 shared benchmarks.
Coding and Software Engineer
Qwen3.8-Max 1/1| Benchmark | Qwen3.8-Max | Qwen3.6-Max-Preview | Diff |
|---|---|---|---|
| SWE-Bench Pro - Public | 67.705 / 57极高强度思考（工具） | 57.3020 / 57Deep Thinking (With Tools) | +10.40 |
General Evaluation
Qwen3.8-Max 1/1| Benchmark | Qwen3.8-Max | Qwen3.6-Max-Preview | Diff |
|---|---|---|---|
| GPQA Diamond | 92.6021 / 226极高强度思考（无工具） | 90.4036 / 226最高（无工具） | +2.20 |
General Knowledge
Qwen3.8-Max 1/1| Benchmark | Qwen3.8-Max | Qwen3.6-Max-Preview | Diff |
|---|---|---|---|
| HLE | 56.2013 / 181极高强度思考（工具） | 50.2029 / 181Thinking (With Tools) | +6 |
Specs
| Field | Qwen3.8-Max | Qwen3.6-Max-Preview |
|---|---|---|
| Publisher | 阿里巴巴 | 阿里巴巴 |
| Release date | 2026-08-03 | 2026-04-18 |
| Model type | Reasoning model | Chat model |
| Architecture | MoE | Dense |
| Parameters | 2.4T | Not available |
| Context length | 1M | 262K |
| Max output | 128K | 64K |
API pricing
Prices use DataLearner records when available; missing fields are not inferred.
| Item | Qwen3.8-Max | Qwen3.6-Max-Preview |
|---|---|---|
| Text input | ¥12 / 1M tokens | $1.3 / 1M tokens |
| Text output | ¥36 / 1M tokens | $7.8 / 1M tokens |
| Cache read | ¥1.5 / 1M tokens | Not public |
Summary
- Qwen3.8-Maxleads in:Coding and Software Engineer (1/1), General Evaluation (1/1), General Knowledge (1/1)
On average across the 3 shared benchmarks, Qwen3.8-Max scores 6.20 higher.
Largest single-benchmark gap: SWE-Bench Pro - Public — Qwen3.8-Max 67.70 vs Qwen3.6-Max-Preview 57.30 (+10.40).
Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.