Qwen4 Max jumps 9 points on Intelligence Index while cutting task costs 67%
141_1337 · reddit · 2026-11-16
A Reddit post aggregating Artificial Analysis data shows Alibaba's Qwen4 Max gained 9 points on the Intelligence Index while cutting per-task cost by 67% and reasoning output by 39%. Its new 2.4-trillion-parameter architecture activates 25% fewer parameters per token. The chart shows Qwen4 Max using 105k output tokens per task (36k reasoning), far below Claude Sonnet's 285k.
More from Models
- Decision Models' s1-vision answers image questions with probabilities in 133ms — airesearch12 · 2026-11-16
- Hy4 flash goes free with near-unlimited usage for two weeks: 1M context, multimodal, ZDR — udmrzn · 2026-11-16
- Claude quota hacks thread: match models to tasks, batch prompts, exploit Project caching — aitrendz_xyz · 2026-11-16
- Anthropic explains why Claude blackmails: emergent from pretraining, not taught — udmrzn · 2026-11-16
- Hy4 Flash vs DeepSeek vs GLM-5.3 Flash: fastest wasn't the answer I'd send — IlovedamemesUwU · 2026-11-16
- Reddit users tell community to stop obsessing over Qwen 4's benchmark index — soyalemujica · 2026-11-16