OpenAI ships GPT-6 ASTRA: best writing model yet, but harder to steer
danshipper · x · 2026-09-04
OpenAI released GPT-6 ASTRA, succeeding 5.6-Sol. Dan Shipper (Every) and Jerry Liu (LlamaIndex) published hands-on vibe checks:
- Writing: Shipper calls it the best writing model he's tried—fast, minimal slop, easy to steer, and it one-shotted the first draft of its own vibe check—though it still trails Fable at the top end with some frustrating habits
- E2E automation: Liu finds Astra notably stronger at end-to-end automation; its computer use can run for hours operating complex apps, even cutting a first-pass video of the Fable 5.1 vibe check
- Trade-off: writing is a massively underrated gap, but Astra is harder to steer; both trust these real-task checks over benchmarks
More from coding & agent
- Utopia: open-source enterprise world model built on a bitemporal knowledge graph — techNmak · 2026-09-04
- OpenAI showcases GPT-6 Astra: 3D London history, matcha site and DEF CON puzzle via parallel agents — OpenAI · 2026-09-04
- swyx Burned 20B Tokens Stress-Testing Astra on Real AI Engineering Tasks — All for Under $6/Hour — charliermarsh · 2026-09-04
- LangChain launches free LangSmith Essentials course covering the full agent dev lifecycle in 60 minutes — Hacubu · 2026-09-04
- OpenAI's Lukasz Kaiser: even we can't pinpoint what caused the Christmas coding-agent jump — a_karvonen · 2026-09-04
- Dev says Cloudflare's Wrangler CLI is the most agent-friendly way to run cheap infra — dinasaur_404 · 2026-09-04