Qwen: Qwen3.8 Omni Flash

Qwen3.8 Omni Flash, an omni-modal reasoning model from Alibaba, demonstrates a mixed performance profile with notable strengths in specific areas. The model tends to have longer response times, ranking in the 11th percentile for speed across benchmarks. However, it offers competitive pricing, placing in the 43rd percentile. Reliability is a significant strong point, with a 95% success rate indicating consistent and usable responses. In terms of accuracy, Qwen3.8 Omni Flash excels in Email Classification (99.0% accuracy, 86th percentile) and shows strong capabilities in Instruction Following (79.0% accuracy, 81st percentile). Its General Knowledge (96.0%) and Coding (86.9%) performances are solid, though not top-tier. A key weakness is observed in Reasoning, where it achieved only 56.0% accuracy (28th percentile), and in its ability to acknowledge uncertainty, with an 86.0% accuracy in Hallucinations (27th percentile), suggesting it may not always appropriately indicate when it lacks knowledge. Despite its agentic capabilities and native audio-video understanding, its overall speed remains a challenge.