Sapiens AI

Agnes 2.5 Pro Alpha (based on Qwen3.5 397B A17B)

Jul 24, 2026

Compare

Intelligence

26.8

#158 of 751Rank 158 of 751 measured models

Coding

58.8

#89 of 298Rank 89 of 298 measured models

Output speed

160 t/s

#83 of 411Rank 83 of 411 measured models

Price/1M

$0.563

#314 of 677Rank 314 of 677 measured models

Context window

1M

#89 of 1256Rank 89 of 1256 measured models

Agnes 2.5 Pro Alpha (based on Qwen3.5 397B A17B) scores 26.8 on the Artificial Analysis Intelligence Index, ranking 158th of 751 measured models. The closest models on this index are Kimi K2.6 (27.0), Qwen3.6 Plus (27.0), and Quasar 438B (Max, Based on GLM-5.2) (26.7). At $0.563 per million tokens (3:1 input/output blended price), it is cheaper than Kimi K2.6 ($1.71) and Qwen3.6 Plus ($1.13), more expensive than Quasar 438B (Max, Based on GLM-5.2) ($0.165). With an output speed of 160 t/s, it is faster than Kimi K2.6 (59.4 t/s) and Qwen3.6 Plus (55.9 t/s), slower than Quasar 438B (Max, Based on GLM-5.2) (187 t/s).