GPT-6 Sol and Luna cut costs in half while matching GPT-5.6 scores
As seen on the 24/7 Wall St. homepage on September 22, 2026.
A new flagship that costs half as much without scoring higher says the money in AI is shifting from capability to cost per token, and inference margins across the model vendors get squeezed with it.
GPT-6 Sol and Luna push the cost efficiency frontier by halving cost relative to GPT-5.6 Sol and Luna. Intelligence Index and Coding Agent Index scores remain level with GPT-5.6, with progress in some evaluations and regressions in others Pricing is approximately half that of https://t.co/eubnxPVsyN
- Replies25
- Reposts20
- Likes201
Continue ReadingShow less
GPT-6 Sol and Luna arrive priced at approximately half what GPT-5.6 Sol and Luna cost, and that is essentially the whole story on capability: Artificial Analysis finds the Intelligence Index and Coding Agent Index scores sitting level with the prior generation, with some evaluations improving and others regressing.
That combination, a flagship price cut without a meaningful benchmark leap, signals where the competitive pressure in AI is moving. The race now runs on cost per token, and OpenAI just reset the floor with this release.
Sponsored
Twelve Tabs, One Thesis
Your Research Resets Every Morning
The quote page in one tab. Filings in another. A chart you rebuilt from scratch, a transcript you never went back and found, a screener whose settings you will redo next week. Nothing you built yesterday is still there.
AlphaSpace replaces all of it with one screen you arrange yourself. Earnings calendar, estimate versus actual, the call transcript, live news, your own charts, every panel wired to whatever ticker you click. Close the browser and it is all still sitting there tomorrow.
See What a Built View Looks Like →
(Sponsor)
For investors watching the broader AI infrastructure trade, the implications run downstream. When the leading model provider halves its pricing without adding headroom on performance, every other model vendor faces the same arithmetic: compress margins or lose share. Inference economics across the industry tighten with each move like this.
The post from Artificial Analysis drew more than 13,000 impressions within hours of going live, showing how strongly the cost efficiency framing registered with practitioners and market observers alike. The benchmark parity detail, progress in some evaluations offset by regressions in others, is the number to watch as third-party testing firms run fuller evaluations in the days ahead.