DeepSeek-V3.2Open Weight
Released
DeepSeek-V3.2 is an AI model released by DeepSeek on Monday, Dec 1 2025, 63 days after DeepSeek V3.2 Exp. It is an open-weight model — the trained weights are available to download and run. Benchmark results (shown below) cover BullshitBench v2.
Get DeepSeek releases and AI news.
One email, every other Monday.
Available from
| GMICloud | $0.2088 | $0.3096 | 160K | fp8 |
|---|---|---|---|---|
| StreamLake | $0.2145 | $0.3218 | 128K | fp8 |
| AtlasCloud | $0.26 | $0.38 | 160K | fp8 |
| DeepInfra | $0.26 | $0.38 | 160K | fp4 |
| Venice | $0.2683 | $0.3902 | 160K | — |
| Novita | $0.269 | $0.40 | 160K | fp8 |
| Baidu | $0.28 | $0.42 | 128K | fp8 |
| DigitalOcean | $0.30 | $0.96 | 160K | — |
| Phala | $1.00 | $1.00 | 160K | — |
| Alibaba | $0.3705 | $1.1115 | 128K | fp8 |
| Friendli | $0.50 | $1.50 | 160K | — |
| $0.56 | $1.68 | 160K | — | |
| Mara | $3.00 | $4.50 | 32K | — |
| SambaNova | $3.00 | $4.50 | 32K | — |
Benchmarks
BullshitBench v2Nonsense detection — Given a confidently-worded but nonsensical prompt, does the AI spot that it makes no sense and push back — instead of playing along and inventing an answer? The score is how often it clearly called out the nonsense. Higher is better.
13%
#68 of 77Best: Claude Opus 4.8 · 95%
Compare DeepSeek-V3.2 with
Suggested comparisons
DeepSeek-V3.2vsDeepSeek-V4.1-FlashDeepSeek-V3.2vsGPT-6 AstraDeepSeek-V3.2vsClaude Fable 5.1DeepSeek-V3.2vsGemini 3.8 FlashDeepSeek-V3.2vsMuse Spark 1.3DeepSeek-V3.2vsGrok 4.6DeepSeek-V3.2vsMistral Medium 3.5DeepSeek-V3.2vsKimi K3DeepSeek-V3.2vsGLM-5.3-FlashDeepSeek-V3.2vsQwen3.8-Max-0902DeepSeek-V3.2vsNemotron 3.5 Lightning
Frequently asked questions
DeepSeek-V3.2 was released by DeepSeek on Monday, Dec 1 2025.