Compare / head-to-head
Qwen3.5 397B-A17BvsQwen3.8-Max
Qwen3.8-Max leads 4 of 4 shared benchmarks. Qwen3.5 397B-A17B is 5.5x cheaper per token. Qwen3.8-Max has the larger context window (1M tokens).
Benchmarks from cited public sources; pricing from official pages; status from official provider feeds.
Shared benchmarks
0 – 4
Qwen3.8-Max leads
Cheaper per token
Qwen3.5 397B-A17B
5.5x cheaper, input + output
Larger context
Qwen3.8-Max
1M tokens
Provider uptime (30d)
—
Qwen
Qwen3.5 397B-A17B
Qwen · released 2026-02-15
textvision
Qwen3.8-Max
Qwen · released 2026-08-02
textvision
Quality
Benchmark matrix
4 shared · 3 only Qwen3.5 397B-A17B · 4 only Qwen3.8-Max| Benchmark | Qwen3.5 397B-A17B | Qwen3.8-Max | Δ | Edge |
|---|---|---|---|---|
| Reported by both · 4 | ||||
| GPQA Diamond | 88.4% ↗ | 92.6% ↗ | -4.2 pt | Qwen3.8-Max |
| LMArena Elo | 1438.3 ↗ | 1480.6 ↗ | -42.3 | Qwen3.8-Max |
| FrontierMath Tiers 1-3 v2 (Epoch AI run) | 29.5% ↗epoch | 74.7% ↗epoch | -45.2 pt | Qwen3.8-Max |
| OTIS Mock AIME 2024-2025 (Epoch AI run) | 88.9% ↗epoch | 99.4% ↗epoch | -10.5 pt | Qwen3.8-Max |
| Only Qwen3.5 397B-A17B reports · 3 | ||||
| SWE-bench Verified | 76.4% ↗ | not reported | — | — |
| MMLU-Pro | 87.8% ↗ | not reported | — | — |
| AIME 2026 | 94.2% ↗matharena ⚠ | not reported | — | — |
| Only Qwen3.8-Max reports · 4 | ||||
| FrontierMath Tier 4 v2 (Epoch AI run) | not reported | 46.3% ↗epoch | — | — |
| SimpleQA Verified | not reported | 45.8% ↗epoch | — | — |
| SWE-Bench Pro | not reported | 67.7% ↗ | — | — |
| Terminal-Bench 2.1 | not reported | 86.6% ↗ | — | — |
Scores tagged "epoch" or "matharena" are independent runs, used only where the lab has not published its own; ⚠ marks rows MathArena flags as released after the competition. Higher is better on every row. Δ is Qwen3.5 397B-A17B minus Qwen3.8-Max in the benchmark's own unit. "Not reported" means the lab has not published that figure; it is not a zero. ↗ opens the source.
Specs & pricing
Side by side
Official pricing pages and model cards| Spec | Qwen3.5 397B-A17B | Qwen3.8-Max | Edge |
|---|---|---|---|
| Context window | 262K tokens | 1M tokens | Qwen3.8-Max |
| Max output | — | 131K tokens | — |
| Input price / 1M | $0.172 ↗ | $1.65 ↗ | Qwen3.5 397B-A17B |
| Output price / 1M | $1.03 ↗ | $4.95 ↗ | Qwen3.5 397B-A17B |
| Cached input / 1M | — | — | — |
| Input + output / 1M Lower is cheaper. List prices; batch, tool and regional fees excluded. | $1.2 | $6.6 | Qwen3.5 397B-A17B |
| Modalities | text · vision | text · vision | Tie |
| Released | 2026-02-15 | 2026-08-02 | — |
| Cited benchmark scores | 8 | 9 | — |
Reliability
Qwen status
Last 60 days · refreshed daily on this page · live on /statusMore matchups
Qwen3.5 397B-A17B vs …
Models sharing the most benchmarksMore matchups
Qwen3.8-Max vs …
Models sharing the most benchmarksBuilt by Respan
Which one wins on your data?
Public benchmarks are a starting point. Run Qwen3.5 397B-A17B and Qwen3.8-Max on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.