A model priced 80% below the frontier reframes the comparison

Claim

At $2 and $6 per million tokens, the model's output costs 80% less than GPT-5.6 Sol and 88% less than Fable 5, which is why a model that works for ten days beats a better model you can afford to run for ten minutes, lifting Alibaba up to 7.3% in Hong Kong.

Reported metric: -80% / -88%

Reporting date: 2026-08-03 (not necessarily the event date)

Sources

Attribution and reporting linked by this record; links do not imply independent verification.

• Welcome to August 3, 2026 (theinnermostloop.substack.com)