DeepSeek-V3.2 (Thinking): benchmarks, price and rank
DeepSeek-V3.2 (Thinking) from DeepSeek is 98th of 123 rated AI models on Spezia's leaderboard. It costs $0.28 in, $0.42 out per million tokens.
- 98th of 123 rated models
- Rating 799
- Rated on 6 benchmarks
- Open weights
- Released Dec 1, 2025
- 131,000 tokens of context
Where it is strongest
Not in the top tenth on any benchmark in use.
Where it falls behind
- 33rd of 37 on Toolathlon
- 44th of 56 on BrowseComp
- 79th of 103 on GPQA Diamond
Price: $0.28 in, $0.42 out per million tokens.
Its headline results
The benchmarks that count most in the rating, with its place among the models measured the same way, and where each number came from.
- BrowseComp51.4%
Questions whose answers are hidden deep on the web and take many searches and pages to track down.
44th of 56. The maker's own number, from DeepSeek, via LLM Stats.
- Humanity's Last Exam25.1%
About 2,500 questions written by experts in more than a hundred fields to be as hard as they could make them, many needing a diagram or image.
46th of 74. The maker's own number, from DeepSeek, via LLM Stats.
- Toolathlon35.2%
Long tasks across many real apps and their tools, such as email, calendars, spreadsheets, code repositories and online stores.
33rd of 37. The maker's own number, from DeepSeek, via LLM Stats.
- SWE-bench Verified73.1%
Real bug reports and feature requests from popular open-source Python projects. The model has to change the code so the project's hidden tests pass.
23rd of 56. The maker's own number, from DeepSeek, via LLM Stats.
- GPQA Diamond83.4%
Graduate-level questions in biology, physics and chemistry, written by PhDs so that searching the web does not give the answer away.
79th of 103. Independent result from Epoch AI.
- MMLU-Pro85%
Twelve thousand multiple-choice questions across fourteen subjects, from law and medicine to engineering, with ten options each.
18th of 56. The maker's own number, from DeepSeek, via LLM Stats.
1 more results in its full profile.
Every result with its source, what it costs per task, what jumped with its release, side by side with any other model, and answers to your questions about when to use it.
Models ranked near it
How the rating works, and every benchmark it uses: the model leaderboard.