TAU2
Measured May 14, 2026Source
Score
0.85
Models9 1 
Qwen3.6 35B A3B (Non-reasoning)
阿里巴巴阿里巴巴China
This is a 35-billion parameter model from Alibaba's Qwen series, optimized for fast and efficient inference without the overhead of a dedicated reasoning mode. It maintains strong capabilities in coding and multimodal tasks while offering a balance between performance and cost-effectiveness.
CodingMultimodalFastCheap
Input / 1M tokens
$0.375Output / 1M tokens
$2.25Output tokens/s
183.3First-token seconds
1.4sSupported plans
4Benchmark history
Evaluations
Terminalbench Hard
Measured May 14, 2026Source
Score
0.26
Lcr
Measured May 14, 2026Source
Score
0.57
Ifbench
Measured May 14, 2026Source
Score
0.36
Scicode
Measured May 14, 2026Source
Score
0.01
Hle
Measured May 14, 2026Source
Score
0.13
Gpqa
Measured May 14, 2026Source
Score
0.82
Artificial Analysis Coding Index
Measured May 14, 2026Source
Score
17.6
Artificial Analysis Intelligence Index
Measured May 14, 2026Source
Score
31.5
Plan availability
Products and plans that support this model
Neuralwatt
Neuralwatt
Neuralwatt Cloud is an AI inference API with energy-based pricing, providing OpenAI-compatible access to multiple models from providers like Mistral, ZhipuAI, OpenAI, Kimi, MiniMax, and Qwen, with real-time energy tracking.
Discussion
Thinking... Make sure you are connected to GitHub server