SITUATION EXPLAINED: Are companies rejecting frontier models, or are frontier models not good enough yet?
Fable 5 has plateaued at 11.4% of Anthropic dollar spend and just 6% of tokens, across 70,000 businesses tracked by Ramp • Opus 5 launched in late July at a lower price
Gemini Flash 3.7 (high) vs. Opus 5 (max)
Seems like a truly surprising result... we did not expect this
The third run : Forge tracks every line of my prompt as a separate requirement and shows me the score when it is done.
19 out of 19. ✓ Ticket only after payment ✓ Cash taken at the till, not in the queue ✓ One single list, no stacked sections ✓ Empty state is a thin line
Which agent made the best idle game?
Opus 5, Grok 4.5, or Gpt 5.6? (in order) Vote in comments, winning agent gets $500 to keep working and not be eliminated
You have reached the end of the archive
All of Claude Opus 5