Model card
DeepSeek V3.5.
DeepSeekopen-source685B MoE params
§ 01 · Benchmarks
Every benchmark DeepSeek V3.5 has a recorded score for.
| # | Benchmark | Area · Task | Metric | Value | Rank | Date | Source |
|---|---|---|---|---|---|---|---|
| 01 | SWE-Bench | Computer Code · Code Generation | resolve-rate-agentic | 74.6% | #14 | 2025-11-01 | |
| 02 | SWE-Bench | Computer Code · Code Generation | resolve-rate | 74.6% | #16 | 2025-11-01 | source ↗ |
Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.
§ 03 · Papers
1 paper with results for DeepSeek V3.5.
- 2023-10-10· Computer Code· 1 result
SWE-bench: Can Language Models Resolve Real-World GitHub Issues?
Carlos E. Jimenez, John Yang, Alexander Wettig, Shunyu Yao et al.
§ 04 · Related models
Other DeepSeek models scored on Codesota.
§ 05 · Sources & freshness
Where these numbers come from.
deepseek-agent
1
result
swebench-leaderboard
1
result
1 of 2 rows marked verified.