Fable 5.2 and the Slowdown
Anthropic released Claude Fable 5.2 today, with Mythos 5.2 alongside it for vetted organisations only. On Friday I called the next Fable for the last week of November, working from the 84 days between Fable 5 and 5.1, and it came after 69. I’d also been calling it Fable 5.5, because that’s the name the leaks used, so I’ve missed on both the date and the number.
CNBC points out that it’s the first frontier release since Dario Amodei called in September for the pace of AI development to slow. Anthropic says it built new evaluations and post-launch monitoring before shipping, and I believe that. The model itself goes the other way, though. Anthropic’s Fable page lists $8 per million input tokens and $40 per million output, down from $10 and $50, and CNBC says it produces output 30% faster. Whatever slowing down means at Anthropic, it doesn’t mean a dearer or slower flagship. With the IPO possibly as soon as next week, a price cut is the obvious way to show investors usage climbing, and I suspect that weighed more than the September call did.
The announcement still tells people to start with Opus 5.5 for most coding and knowledge work, and to save Fable for the hard problems. Fable 5.2 was in Claude Code’s model menu this evening, and I picked it, then went straight back to Opus 5.5.
Sources:
-
Claude Fable — Anthropic
-
Claude Fable 5.2 and Claude Mythos 5.2 — Anthropic
-
Anthropic releases Claude Fable 5.2 and Claude Mythos 5.2 ahead of IPO — CNBC
Filed under AI & machine learning
This post is timestamped using Blockchain technology. Verify