Wednesday, June 10
TL;DR Anthropic shipped Claude Fable 5 and Mythos 5 yesterday, the biggest model release of the year so far. Apple’s WWDC brought Claude into iPhone workflows for the first time via the new Foundation Models Extensions system.
Models & research
-
Anthropic launched model-releases Claude Fable 5 and Claude Mythos 5 on June 9. Fable 5 is the same underlying Mythos-class model with safety classifiers attached; Mythos 5 is the raw version, restricted to vetted cyberdefense partners. Fable 5 hits 80.3% on SWE-Bench Pro vs 69.2% for Opus 4.8 and 58.6% for GPT-5.5. Priced at $10/$50 per million tokens; free for Pro/Max/Team/Enterprise through June 22. Anthropic announcement · TechCrunch
-
karpathy called Fable 5 “a major-version-bump-deserving step change forward (imo of the same order as Claude 4.5 was in November)” and noted it handles very ambitious long-horizon tasks without hand-holding. tweet
People
- simonw posted hands-on impressions of Fable 5 on launch day: put it to work on a gnarly pause-resume problem in tool calls, and it not only solved the problem but found and fixed four issues in his underlying LLM library unprompted. post
Tools & dev
-
At WWDC 2026 (June 8-9), Apple announced a multi-AI Extensions system for the Foundation Models framework that lets developers swap in third-party cloud models including Claude, with no session-logic changes required. Anthropic shipped a Swift package to wire Claude into the
LanguageModelSessionAPI on iOS 27, iPadOS 27, and macOS 27. Anthropic blog · TechCrunch WWDC roundup -
Apple also announced Gemini-powered Siri and iOS 27 Beta 1 at the same event. The multi-AI Extensions system means devs can prototype on-device with Apple’s model and route harder queries to Claude or Gemini by swapping a Swift Package Manager dependency. TechCrunch