DeepSeek V3.2 Exp
DeepSeek / DeepSeek
20scores
19benchmarks
$0.27 / $0.41 per 1M tokenscost in/out
Metadata
DeepSeek Open source
Aliases: deepseek-deepseek-v3.2-exp, deepseek-v3.2-exp, deepseek/deepseek-v3.2-exp
| Benchmark | Category | Rank | Score | Sampled |
|---|---|---|---|---|
| Berkeley Function-Calling Leaderboard | Agentic | 14 | 56.73% | 2026-05-27 |
| Berkeley Function-Calling Leaderboard | Agentic | 19 | 54.12% | 2026-05-27 |
| MCP-Universe | Agentic | 20 | 19.91 | 2026-05-06 |
| Tau2-Bench Telecom | Agentic | 224 | 33.9% | 2026-05-11 |
| Terminal-Bench Hard | Agentic | 91 | 31.1% | 2026-05-11 |
| UAVBench | Agentic | 17 | 73.55 | 2026-05-06 |
| Arena AI Code | Coding | 57 | 1286 | 2026-05-06 |
| Codeforces | Coding | 11 | 0.707 | 2026-05-28 |
| SciCode | Coding | 140 | 37.7% | 2026-05-11 |
| Artificial Analysis Intelligence Index | Intelligence | 124 | 32.94 | 2026-05-11 |
| Humanity's Last Exam | Intelligence | 103 | 13.8% | 2026-05-11 |
| MMLU-Pro | Intelligence | 32 | 85% | 2026-05-11 |
| LEXam | Legal | 7 | 57.42% open / 53.07% MCQ | 2026-05-28 |
| ConStory-Bench | Long Context | 8 | CED 0.541 | 2026-05-28 |
| AIME 2025 | Math | 40 | 87.7% | 2026-05-11 |
| HMMT 2025 | Mathematics | 22 | 0.84 | 2026-05-06 |
| Artificial Analysis Openness Index | Openness | 78 | 44.44 | 2026-05-11 |
| GPQA Diamond | Reasoning | 104 | 79.7% | 2026-05-11 |
| CritPt | Science | 67 | 1.4% | 2026-05-11 |
| BrowseComp-zh | Search | 12 | 0.48 | 2026-05-06 |
No matching rows.