TAU2
Measured May 29, 2026Source
Score
0.91
Models9 1 
DeepSeek V4 Pro (Non-reasoning)
DeepSeekDeepSeekChina
This is the non-reasoning variant of the DeepSeek V4 Pro model. It retains the core capabilities of the V4 Pro series, such as strong coding and general knowledge, but is optimized for faster response times and lower cost by omitting the extended reasoning or chain-of-thought process. It is suitable for applications requiring quick, cost-effective responses.
CodingFastCheapMultimodal
Input / 1M tokens
$0.435Output / 1M tokens
$0.87Output tokens/s
49.48First-token seconds
1.27sSupported plans
6Benchmark history
Evaluations
Terminalbench Hard
Measured May 29, 2026Source
Score
0.36
Lcr
Measured May 29, 2026Source
Score
0.45
Ifbench
Measured May 29, 2026Source
Score
0.46
Scicode
Measured May 29, 2026Source
Score
0.42
Hle
Measured May 29, 2026Source
Score
0.08
Gpqa
Measured May 29, 2026Source
Score
0.72
Artificial Analysis Coding Index
Measured May 29, 2026Source
Score
38.4
Artificial Analysis Intelligence Index
Measured May 29, 2026Source
Score
39.3
Plan availability
Products and plans that support this model
Crof AI Plan
Crof AI Plan
CrofAI is a cost-effective AI inference API platform that provides access to various open-source large language models (LLMs) such as DeepSeek, GLM, and Qwen through an OpenAI-compatible API. It offers tiered subscription plans and pay-as-you-go pricing for developers and businesses.
User ratings
Loading ratings...
Discussion
Thinking... Make sure you are connected to GitHub server