Alibaba

Qwen3.8 Max

Aug 3, 2026ReasoningClosed weights

Compare

Intelligence

40.2

#51 of 763Rank 51 of 763 measured models

Coding

71.8

#44 of 299Rank 44 of 299 measured models

Output speed

38.8 t/s

#395 of 421Rank 395 of 421 measured models

Price/1M

$3

#531 of 718Rank 531 of 718 measured models

Context window

1M

#88 of 1285Rank 88 of 1285 measured models

Cost per task

$2.67

#178 of 215Rank 178 of 215 measured models

Qwen3.8 Max scores 40.2 on the Artificial Analysis Intelligence Index, ranking 51st of 763 measured models. The closest models on this index are Claude Opus 4.7 (40.7), Qwen3.8 2.4T A95B (39.9), and Qwen3.8-Flash-Next (39.8). At $3 per million tokens (3:1 input/output blended price), it is cheaper than Claude Opus 4.7 ($10), priced close to Qwen3.8 2.4T A95B ($3), more expensive than Qwen3.8-Flash-Next ($0.23). With an output speed of 38.8 t/s, it is about as fast as Qwen3.8 2.4T A95B (39 t/s), slower than Claude Opus 4.7 (53.9 t/s) and Qwen3.8-Flash-Next (51.3 t/s).