Loading…
OpenAI's GPT-6 Sol and Luna release signals a claimed improvement in the cost-performance of hosted professional-work models, but the supplied evidence is company-reported.
Limits of the evidence: Benchmark comparisons and factuality results are reported by OpenAI; the factuality evaluation is internal, and no independent corroboration is included.
What to watch
Assessment revised Oct 4, 2026
No sources are attached to this story yet.
First assessment, from OpenAI News
No maturity ladder applies to this desk.
Sens.ai aggregates and assesses published reporting. The assessment above is machine generated; the original sources are authoritative.