Tools ·
GPT-5.3-Codex: OpenAI's agentic coder gets about 25% faster and steerable mid-task
OpenAI released GPT-5.3-Codex, which it describes as its most capable agentic coding model, running about 25% faster than its predecessor and supporting mid-task steering. It reported 56.8% on SWE-Bench Pro and 77.3% on Terminal-Bench 2.0, and was available in paid ChatGPT plans via the Codex app, CLI and IDE extension, with API access to follow. For engineering teams, it raised the baseline for long-running coding agents that can be redirected without losing context.
Read the full story at OpenAI ↗Like, save or discuss on Buzz
