GPT-6-Astra leaps across benchmarks, called 'probably AGI', timed to undercut Anthropic IPO
kimmonismus · x · 2026-09-04
- Based on OpenAI's official presentation (untested by the author), the new Astra base model posts major benchmark gains across browser/computer use, Excel, coding, drug discovery, and math research.
- Altman and Brockman have called it "probably AGI"; the author argues it genuinely qualifies as general intelligence with expert-level performance across domains.
- The timing—days after rival Fable 5.1 and just before Anthropic's planned IPO—is read as a deliberate move to dent Anthropic's IPO appeal ahead of OpenAI's own.
- Another model, codenamed "Bel", reportedly arrives by end of 2026; the author says 2026 is the turning point as AI enters every workplace.
Related event: OpenAI launches GPT-6 Astra, declaring "Welcome to the AGI era"(156 posts)→
More from Models
- Codex Down for Much of Rollout Day — and Users Not Getting New Model Either — RexDouglass · 2026-09-04
- Neuralese explained: why OpenAI's Astra architecture has safety researchers alarmed — ShakeelHashim · 2026-09-04
- K2 Horizon launches six fully open models from 0.9B to 375B with training code, data recipes and logs — aliscodes · 2026-09-04
- Is Astra AGI? Five contradictory answers that are all true at once — shaunralston · 2026-09-04
- Early Access Tester: GPT-6 Astra Built a Blender Werewolf in 8 Minutes and a World Simulator in 17 — TheMoonMidas · 2026-09-04
- swyx Burned 20B Tokens Stress-Testing Astra on Real AI Engineering Tasks — All for Under $6/Hour — charliermarsh · 2026-09-04