← All comparisons

Hy3-preview (Non-reasoning) vs Claude 4.1 Opus (Reasoning)

Tencent vs Anthropic — side-by-side benchmark comparison

Hy3-preview (Non-reasoning)Claude 4.1 Opus (Reasoning)
Intelligence Index33.742.0
Coding Index34.336.5
Math Index80.3
Output speed (tok/s)94.544.5
Blended price ($/1M)$0.20$32.81
Time to first token (s)2.41s8.55s
aime
aime 2580.3%
artificial analysis coding index34.3036.50
artificial analysis intelligence index33.7042.00
artificial analysis math index80.30
gpqa73.2%80.9%
hle6.3%11.9%
ifbench48.0%55.4%
lcr33.7%66.3%
livecodebench65.4%
math 500
mmlu pro88.0%
scicode39.4%40.9%
tau267.5%71.4%
terminalbench hard31.8%34.3%

Benchmark data from Artificial Analysis.