TAU2
Measured May 14, 2026Source
Score
0.34
Models13 4 0 
DeepSeek V3.2 Exp (Reasoning)
DeepSeekDeepSeekChina
An experimental reasoning-focused variant of the DeepSeek V3.2 model, optimized for complex logical deduction and chain-of-thought tasks. It likely maintains strong coding and general capabilities while enhancing performance on problems requiring multi-step reasoning.
ReasoningCodingFastCheap
Input / 1M tokens
$0.275Output / 1M tokens
$0.415Supported plans
0Benchmark history
Evaluations
Terminalbench Hard
Measured May 14, 2026Source
Score
0.31
Lcr
Measured May 14, 2026Source
Score
0.69
Ifbench
Measured May 14, 2026Source
Score
0.54
Aime 25
Measured May 14, 2026Source
Score
0.88
Scicode
Measured May 14, 2026Source
Score
0.38
Livecodebench
Measured May 14, 2026Source
Score
0.79
Hle
Measured May 14, 2026Source
Score
0.14
Gpqa
Measured May 14, 2026Source
Score
0.8
Mmlu Pro
Measured May 14, 2026Source
Score
0.85
Artificial Analysis Math Index
Measured May 14, 2026Source
Score
87.7
Artificial Analysis Coding Index
Measured May 14, 2026Source
Score
33.3
Artificial Analysis Intelligence Index
Measured May 14, 2026Source
Score
32.9
CODPL speed
Provider ranking
61.91
733 ms
100%
3
1h
Rank #1
41.26
962 ms
100%
3
1h
Rank #2
Ollama Cloud
Product: Ollama Cloud45.43
1,501 ms
100%
2
1h
Rank #3
百度云(千帆)
Product: 百度千帆 Coding Plan32.15
1,030 ms
100%
3
1h
Rank #4
Plan availability
Products and plans that support this model
No products or plans have been linked to this model yet.
Discussion
Thinking... Make sure you are connected to GitHub server