← All comparisons

Hy3-preview (Non-reasoning) vs Claude 3.7 Sonnet (Non-reasoning)

Tencent vs Anthropic — side-by-side benchmark comparison

Hy3-preview (Non-reasoning)Claude 3.7 Sonnet (Non-reasoning)
Intelligence Index33.730.8
Coding Index34.326.7
Math Index21.0
Output speed (tok/s)94.50.0
Blended price ($/1M)$0.20$6.56
Time to first token (s)2.41s0.00s
aime22.3%
aime 2521.0%
artificial analysis coding index34.3026.70
artificial analysis intelligence index33.7030.80
artificial analysis math index21.00
gpqa73.2%65.6%
hle6.3%4.8%
ifbench48.0%44.0%
lcr33.7%48.3%
livecodebench39.4%
math 50085.0%
mmlu pro80.3%
scicode39.4%37.6%
tau267.5%50.0%
terminalbench hard31.8%21.2%

Benchmark data from Artificial Analysis.