Models / Compare /DeepSeek V3.2 vs Qwen3.5 397B A17B
Head to head
DeepSeek V3.2 vs Qwen3.5 397B A17B
Compared on live catalog data — pricing synced daily, benchmark scores with provenance.
The short version
- DeepSeek V3.2 is cheaper to run hosted: $0.40/M output vs $2.34/M for Qwen3.5 397B A17B (5.8× difference at the cheapest tracked provider).
- Qwen3.5 397B A17B takes 262K tokens of context vs 164K for DeepSeek V3.2.
- Qwen3.5 397B A17B scores 1490.8235 on Arena Coding vs 1469.5929 for DeepSeek V3.2.
- Qwen3.5 397B A17B scores 1406.7816 on Arena Creative Writing vs 1400.3738 for DeepSeek V3.2.
| Spec | ||
|---|---|---|
| Vendor | DeepSeek | Qwen |
| Parameters | 685B / 37B | 403B / 17B |
| 164K | 262K | |
| Weights | OPEN | OPEN |
| License | MIT License | Apache License 2.0 |
| Modality | text->text | text+image+video->text |
| Released | Dec 1, 2025 | Feb 16, 2026 |
| 432.1 GBest | 254.3 GBest | |
| Cheapest hosted | $0.27 / $0.40 per 1M (Novita AI) | $0.39 / $2.34 per 1M (OpenRouter) |
| Tracked providers | 19 | 15 |
DeepSeek V3.2 — our take
DeepSeek V3.2 is a large downloadable text model with a permissive MIT licence and a 163,840-token request limit. Only 37 billion of its 685.4 billion total parameters are active per token, keeping inference costs down while measured coding performance is the clear peak in its benchmark profile.
full breakdown →Qwen3.5 397B A17B — our take
Qwen 3.5 is a large downloadable model from Alibaba with a permissive Apache licence and a mixture-of-experts design that keeps 17 billion parameters active per token. It scores near the top on measured coding and hard-prompt reasoning among open-weights models, and accepts text, images and video.
full breakdown →Need more columns? Use theinteractive comparison to add up to four models.