Creator
Alibaba
Released
2025-04-28
Intelligence
17.0
Artificial Analysis Index
Coding
14.0
Artificial Analysis Index
In $/1M
$0.45
input tokens
Out $/1M
$1.80
output tokens
Blended $/1M
$0.79
3:1 blended
Speed
66
tokens / sec
Benchmark breakdown
Independent evaluation scores, normalized to 0–100.
GPQA Diamond
61
Humanity’s Last Exam
5
MMLU-Pro
76
SciCode
30
LiveCodeBench
34
MATH-500
90
AIME 2025
24
τ²-Bench (agentic)
27
Terminal-Bench Hard
6
IFBench
37