OpenAI ships a new coding model as agentic programming enters the long-horizon era
OpenAI's latest model targets agentic coding with multi-hour autonomous tasks and a new SWE-bench record, heating up the coding-assistant race.
Industry moves, model releases and product updates · 3 articles · Updated daily
OpenAI's latest model targets agentic coding with multi-hour autonomous tasks and a new SWE-bench record, heating up the coding-assistant race.
Anthropic has published Claude's Skills capability as an open standard, letting developers write reusable skill packs for AI agents. Several platforms have pledged compatibility.
Google's latest image-editing model is trending for its character consistency — one sentence now does what used to require pro software, and rivals are racing to catch up.