Qwen: Qwen3.5-122B-A10B

Qwen3.5-122B-A10B, a native vision-language model with a hybrid architecture, demonstrates exceptional inference efficiency due to its linear attention and sparse mixture-of-experts integration. Created on February 25, 2026, with a context length of 262144, it ranks second only to Qwen3.5-397B-A17B in overall performance. The model consistently performs among the fastest models and offers highly competitive pricing. Its reliability is strong, achieving an 88% success rate across 8 benchmarks. In terms of specific capabilities, Qwen3.5-122B-A10B exhibits a high accuracy of 98.0% in Hallucinations (63rd percentile) and Email Classification (51st percentile), alongside strong instruction following at 75.0% accuracy (80th percentile). Notably, its text capabilities significantly outperform Qwen3-235B-2507, and its visual capabilities surpass Qwen3-VL-235B. However, the model currently shows 0.0% accuracy in General Knowledge, Coding, Reasoning, Ethics, and Mathematics benchmarks, indicating significant areas for improvement in these domains.