GPT-6 Sol leaks, screenshots and release rumors

A comparison chart claims Opus 5.5 beats GPT-6 Sol on intelligence, coding and automation benchmarks, while Sol uses far fewer tokens and costs half as much.

A post with an image asks whether the rumored GPT-6 Sol will come to ChatGPT or remain only in Codex.

A user lists Grok 4.7, Opus 5.5, GPT-6 Sol and Luna as launching on the same day, and notes Gemini 4 is still in development.

A tester says GPT-6 Sol tends to close out a single provable finding while GPT-5.6 Sol keeps probing edge cases, and claims astra=sol and sol=terra are rebadges.

On Next.js evals, GPT-6 Sol, Opus 5.5 and Fable 5.1 reach 97% while Grok 4.7 hits 94%, but Grok is said to deliver more output per dollar.

The author says Opus 5.5 is clearly stronger than Opus 5 while GPT-6 Sol shows limited gains over GPT-5.6 Sol, so they ran 10 real use cases head to head.

A full Japanese prompt for turning a full-body photo into a fisheye-lens subway car ad poster, which the author says should also work on 5.6 Sol.

The post notes that Claude Opus 5.5 and GPT-6 Sol and Luna all appeared yesterday and asks whether the AI race has shifted from the smartest model to delivering the most intelligence at scale.

Third-party platform Aria published per-million-token pricing for GPT-6 Sol, Luna and Claude Opus 5.5, saying it is 73.5% to 94.5% lower than standard direct API rates.

A short image post remarking that the release of GPT-6 Sol and Luna still lands with real force, with no further details.

A video says OpenAI launched GPT-6 Sol and Luna with API prices halved to $2/$10 and $0.10/$0.50 and added usage carryover.

A Japanese user says Sol's intelligence is nearly unchanged from 5.6 with API at half price, Luna is far cheaper, but Opus 5.5 was the real star of the day.