OpenAI Launches GPT-Live
AI寒武纪 · wechat · 2026-07-09
A post covering OpenAI's release of the new voice model GPT-Live. It features a full-duplex architecture supporting listening while speaking, interruptions, pauses, and more natural back-and-forth conversations. When search or deep reasoning is needed, it offloads the task to a more powerful GPT-5.5 running in the background. The rollout has hit mobile and web clients, with paid users defaulting to GPT-Live-1 and free users to GPT-Live-1mini.
The post also notes that OpenAI audited SWE-BenchPro, concluding it no longer reliably measures frontier coding capabilities, and withdrew it as a leading coding benchmark. OpenAI's review, conducted using model investigators and five senior engineers, determined that coding evaluations need to be harder, fairer, and more trustworthy.
Related event: OpenAI Launches Full-Duplex Voice Model GPT-Live(44 posts)→
More from coding & agent
- 'My intelligence stack': developer splits AI coding work across Opus 5.5, Fable and Astra with human review gates — sh_reya · 2026-10-27
- Baseten webinar: evaluate open models for coding agents by benchmarking on your own merged PRs — baseten · 2026-10-27
- Thorsten Ball on working with agents: run more evals, automate everything, use the best models — omarsar0 · 2026-10-27
- DeepMind paper: LLMs can't sample randomly on their own — pass random numbers from code instead — rohanpaul_ai · 2026-10-27
- BELL: Open-Source Stack Combines Browserbase, Eve, iMessage and Stripe Link for Personal Agents — jeff_weinstein · 2026-10-27
- Legora runs 2,000 parallel agent evals on Modal Sandboxes for legal AI — blelbach · 2026-10-27