ReactBench

A new analysis argues that perceived differences in UI quality between GPT and Claude models stem not from inherent model capability but from restrictive system prompts and hidden instructions. The ar…

A developer discovered that OpenAI's Codex model manifest for GPT-5.5 contained a large frontend guidance block with specific UI rules, suggesting that GPT's perceived poor UI taste may stem from hidd…

ReactBench v1, a new evaluation for coding agents on realistic React work, finds that top models like GPT 5.6 Sol score only 43.1% on tests that measure performance, accessibility, and quality issues …