Qwen3.8 Max: Alibaba's strongest model more often guesses and costs more per task

Alibaba's Qwen3.8 Max achieves 56 points in the Artificial Analysis Intelligence Index, 10 more than Qwen3.7 Max, tying with Claude Opus 4.8. However, it is more expensive per task ($1.14 vs $0.53) and shows increased hallucination rates (40% vs 23%) on the AA-Omniscience test.

Alibaba/Qwen Anthropic Moonshot AI

Zuckerberg Goes All In Against DeepSeek

Mark Zuckerberg's Meta is intensifying its competition with Chinese AI startup DeepSeek. The article highlights Meta's aggressive push in AI, possibly including investments and model development to counter DeepSeek's rise.

DeepSeek Meta