Creator
Alibaba
Released
2026-03-30
Intelligence
38.6
Artificial Analysis Index
Coding
27.6
Artificial Analysis Index
In $/1M
$0.40
input tokens
Out $/1M
$4.80
output tokens
Blended $/1M
$1.50
3:1 blended
Speed
50
tokens / sec
Benchmark breakdown
Independent evaluation scores, normalized to 0–100.
GPQA Diamond
83
Humanity’s Last Exam
14
MMLU-Pro—
SciCode
41
LiveCodeBench—
MATH-500—
AIME 2025—
τ²-Bench (agentic)
88
Terminal-Bench Hard
21
IFBench
51