gpt-6VIDEO

Codex user tests GPT-6.1 Sol: generous usage but very slow

A week on GPT-6.1 Sol at MAX used only 2% of the weekly quota and 2.8M tokens, but the model is extremely slow with mediocre quality, so the author suggests Sol as implementor and Astra as orchestrator.

shownotover

@shownotoverArchived 28 days agoOriginally posted 2026.09.30

shownotover @shownotoverOriginally posted 2026.09.30

Open original on X

🚨 codex users!! We have a good and a bad news!! I used GPT-6.1 Sol at MAX and I got: - 2% weekly usage - 2.8 Million tokens - 52 mins of work The usage is GOOD!! You can use it for a LOT of work for only $20 Now the "BAD" news It is EXTREMELY SLOW!! And quality isn't great... (Check video) The city is smaller than what GPT-6 Astra built!! AND the car animations are lagging... I'd suggest using it as "implementor" and Astra as "orchestrator". NOTE: I used MAX because THAT is only on par with GPT-6 Astra.

x.com/shownotover/status/2105061014001930544Archived 2026.09.30 10:20