Compare / head-to-head
DeepSeek V3.2vs
Qwen3.5 397B-A17B
Qwen3.5 397B-A17B leads 6 of 7 shared benchmarks. Qwen3.5 397B-A17B has the larger context window (262K tokens).
Benchmarks from cited public sources; pricing from official pages; status from official provider feeds.
Shared benchmarks
0 – 6
Qwen3.5 397B-A17B leads · 1 tied
Cheaper per token
—
list price, input + output
Larger context
Qwen3.5 397B-A17B
262K tokens
Providers
2 providers
DeepSeek · Qwen
DeepSeek V3.2
DeepSeek · released 2025-12-01
text
- Context
- 131K
- Max out
- 66K
- Input /1M
- —
- Output /1M
- —
- Cached /1M
- —
- Scores
- 11 · 8 core
Qwen3.5 397B-A17B
Qwen · released 2026-02-15
textvision
Quality
Benchmark matrix
7 shared · 2 only DeepSeek V3.2 · 10 only Qwen3.5 397B-A17B| Benchmark | DeepSeek V3.2 | Qwen3.5 397B-A17B | Δ | Edge |
|---|---|---|---|---|
| Reported by both · 7 | ||||
| SWE-bench Verified | 73.1% ↗ | 76.4% ↗ | -3.3 pt | Qwen3.5 397B-A17B |
| GPQA Diamond | 82.4% ↗ | 88.4% ↗ | -6 pt | Qwen3.5 397B-A17B |
| MMLU-Pro | 85% ↗ | 87.8% ↗ | -2.8 pt | Qwen3.5 397B-A17B |
| AIME 2026 | 94.2% ↗matharena | 94.2% ↗matharena ⚠ | 0 | Tie |
| LMArena Elo | 1424.8 ↗ | 1438.3 ↗ | -13.5 | Qwen3.5 397B-A17B |
| APEX-Agents (Mercor) | 21.3% ↗ | 24.9% ↗ | -3.6 pt | Qwen3.5 397B-A17B |
| LMArena WebDev (LMArena) | 1361.7 ↗ | 1399.4 ↗ | -37.7 | Qwen3.5 397B-A17B |
| Only DeepSeek V3.2 reports · 2 | ||||
| AIME 2025 | 93.1% ↗ | not reported | — | — |
| Humanity's Last Exam (no tools) | 25.1% ↗ | not reported | — | — |
| Only Qwen3.5 397B-A17B reports · 10 | ||||
| Chess Puzzles (Epoch AI run) | not reported | 13% ↗epoch run | — | — |
| FrontierMath Tiers 1-3 v2 (Epoch AI run) | not reported | 29.5% ↗epoch run | — | — |
| LMArena Vision (LMArena) | not reported | 1246.9 ↗ | — | — |
| Mystery Game Puzzles (Epoch AI run) | not reported | 18% ↗epoch run | — | — |
| OTIS Mock AIME 2024-2025 (Epoch AI run) | not reported | 88.9% ↗epoch run | — | — |
| tau2-bench Airline (Sierra) | not reported | 81.5% ↗ | — | — |
| tau2-bench Banking Knowledge (Sierra) | not reported | 9.8% ↗ | — | — |
| tau2-bench Retail (Sierra) | not reported | 84.4% ↗ | — | — |
| tau2-bench Telecom (Sierra) | not reported | 97.8% ↗ | — | — |
| Toolathlon-Verified (HKUST) | not reported | 40.7% ↗ | — | — |
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 DeepSeek V3.2 minus Qwen3.5 397B-A17B 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 | DeepSeek V3.2 | Qwen3.5 397B-A17B | Edge |
|---|---|---|---|
| Context window | 131K tokens | 262K tokens | Qwen3.5 397B-A17B |
| Max output | 66K tokens | — | — |
| Input price / 1M | — | $0.6 ↗ | — |
| Output price / 1M | — | $3.6 ↗ | — |
| Cached input / 1M | — | — | — |
| Input + output / 1M Lower is cheaper. List prices; batch, tool and regional fees excluded. | — | $4.2 | — |
| Modalities | text | text · vision | Qwen3.5 397B-A17B |
| Released | 2025-12-01 | 2026-02-15 | — |
| Cited benchmark scores | 11 | 22 | — |
Reliability
Provider status
Live from /statusMore matchups
DeepSeek V3.2 vs …
Models sharing the most benchmarksMore matchups
Qwen3.5 397B-A17B vs …
Models sharing the most benchmarksBuilt by Respan
Which one wins on your data?
Public benchmarks are a starting point. Run DeepSeek V3.2 and Qwen3.5 397B-A17B on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.