DeepSeek · model passport
DeepSeek-V3.2-Speciale
deepseek-v3-2-speciale- Best percentile
- 99.2LiveCodeBench (Artificial Analysis)
- Composite standing
- 67.8shrunk · 5/27 axes
What the data says
Benchmarked on 5 of 27 tracked axes. Every figure below is the model's own published result on that board.
Strongest axes
- 99.2 pct — LiveCodeBench (Artificial Analysis) (89.6%)
- 95.5 pct — MMLU-Pro (Artificial Analysis) (86.3%)
Uneven: 45.2 percentile points between its best and worst axis — 99.2 pct on LiveCodeBench (Artificial Analysis) versus 54 pct on SimpleBench.
- LiveCodeBench (Artificial Analysis)99.2 pct89.6%Verify →
- MMLU-Pro (Artificial Analysis)95.5 pct86.3%Verify →
- IFBench (Artificial Analysis)72.8 pct63.9%Verify →
- LiveBench70.4 pct64.1%Verify →
- SimpleBench54 pct52.6%Verify →
- OverallNot benchmarked
- WebDev ArenaNot benchmarked
- Vision ArenaNot benchmarked
- Factuality ArenaNot benchmarked
- Search ArenaNot benchmarked
- Document ArenaNot benchmarked
- Agent ArenaNot benchmarked
- GPQA DiamondNot benchmarked
- AIME (math)Not benchmarked
- FrontierMathNot benchmarked
- SimpleQA VerifiedNot benchmarked
- ARC-AGI-2Not benchmarked
- Humanity's Last ExamNot benchmarked
- BFCL v4 (function calling)Not benchmarked
- SWE-bench Verified (live board)Not benchmarked
- Terminal-Bench 4.0Not benchmarked
- HHEM Hallucination (Vectara)Not benchmarked
- SWE-bench VerifiedNot benchmarked
- Terminal-BenchNot benchmarked
- CodingNot benchmarked
- Instruction followingNot benchmarked
- Hard promptsNot benchmarked
Percentiles are the model's standing within each board's own population. Axes marked "not benchmarked" have no published result for this model — that is an absence of data, not a low score.
DeepSeek-V3.2-Speciale stands at 67.8 (raw mean 78.4) — measured on 5 of 27 axes.
Every claim is a computed fact from the current board data, phrased by a low-temperature model call and rejected by a validation gate if any number is not in the source row. How the leaderboards work →