Timeline of Qwen3 235B A22B Instruct 2507

Qwen3 235B A22B Instruct 2507 scores 11.99% on Terminal-Bench

Qwen3 235B A22B Instruct 2507 scores 0% on Terminal-Bench

Qwen3 235B A22B Instruct 2507 scores 53.22% on τ²-bench

Qwen3 235B A22B Instruct 2507 scores 51.22% on IFBench

Qwen3 235B A22B Instruct 2507 scores 41.44% on SciCode

Qwen3 235B A22B Instruct 2507 scores 15.89% on Humanity's Last Exam

Qwen3 235B A22B Instruct 2507 scores 78.99% on GPQA

Qwen3 235B A22B Instruct 2507 scores 12.71 on Artificial Analysis Intelligence Index

Qwen3 235B A22B Instruct 2507: context length changed from 262144 to 256000

OpenRouter lists Qwen3 235B A22B Instruct 2507 at $0.22 in / $0.88 out per 1M tokens