GPT-6 Sol leaks, screenshots and release rumors

A Chinese-language complaint says GPT-6 Sol outputs empty talk and stalls on tasks, only getting moving after the user swears at it.

A user's tests found Opus 5.5 far ahead of Opus 5, while GPT-6 Sol came in weaker than GPT-5.6 Sol.

A Portuguese-language poll asks whether people use GPT-6 Sol or GPT-6 Luna more often.

A post compares Anthropic's and OpenAI's launch coordination, complaining that GPT-6 Sol was announced but is still not live in ChatGPT and that quota resets are slow.

A user shares impressions after half a day with GPT-6 Sol and concludes he probably won't use it much going forward.

A Korean post says GPT-6 Astra's impact was overshadowed by Opus 5.5 and that Sol and Luna fell short of expectations, with more interest now in 6.1 Astra and Sol UltraFast.

After half a day with Claude Opus 5.5, he praises its video generation but says Fable 5.1 and GPT-6 Astra are smarter, while Sol falls far behind.

A promotional post claims GPT-6 Sol, Luna and Opus 5.5 are free on the Experiential Labs platform, with screenshots and a link but no verified source.

Given the same prompt in OpenDesign, GPT-6 Sol and Claude Opus 5.5 each built a 3D ancient Chinese city; both handled rendering and detail well but with clearly different styles, and a full benchmark is coming.

Using the same prompt and each provider's native tools on $20-tier subscriptions, he will compare quality, tokens, usage consumption, time taken and API cost, then say which to pick on a budget.

A post with an image captioned "say what you want about GPT-6 Sol, at least it doesn't do this" compares Sol's behavior with that of other models.

A Japanese user says keeping up with AI releases is pointless, since Astra was just the hot topic before GPT-6 Sol, GPT-6 Luna and Opus 5.5 all arrived.