GPT-6 Astra leaks in benchmark: 279/280 enterprise tasks done, zero hallucination
ryanshrout · x · 2026-09-06
Baxate quotes Signal65's preliminary test results for the rumored GPT-6 Astra (unverified, not officially confirmed):
- At maximum reasoning effort, the model completed 279 of 280 real multi-step enterprise jobs end to end and fabricated nothing on unanswerable retrieval questions.
- Its weighted error count is roughly half that of the previous leaders, Claude Fable 5.1 and Muse Spark 1.3.
- At list API prices (input, cached input, and output), it is also cheaper per correct task than the model it displaces.
The thread's core point: don't evaluate models by $/M tokens alone — "cost per task" or "intelligence per token" is what matters for enterprise deployments.
Related event: Leaked Benchmarks Claim GPT-6 Astra Aces Enterprise Tasks(2 posts)→
More from Models
- Gemini tip: enable agentic video understanding for ~90% lower token usage — patloeber · 2026-10-23
- Weekly AI cybersecurity roundup: PHP2Shell tops millions of attack attempts, 12 incidents ranked by severity — Dan_Jeffries1 · 2026-10-23
- Open-model advocates push back on blames that abliterated LLMs drive vulnerability surge — max_paperclips · 2026-10-23
- Fable usage limits aren't gone—the mobile app just got a usage-reset UI update — imjustnewatai · 2026-10-23
- Claude generates an 11-minute doomer film entirely in code: 'We will not ask twice' — ZeroStateReflex · 2026-10-23
- Gemini somehow can't read its own Personal Instructions, user's toggles don't help — Connect_Reporter_799 · 2026-10-23