Anthropic launches Claude Fable 5.1 and Mythos 5.1
Anthropic just shipped two new frontier models built on the same core engine, but with different levels of freedom. One is for everyone, the other reserved for vetted cybersecurity and life sciences teams under trusted access
This quietly just flipped the entire ai race
Moon dev: "i have been a claude guy this whole time and it kills me to say what i'm about to say" "opus 5 versus gpt 5.6 and the number one spot is not who you think it is anymore" "the math is not even close and once you see it you
Gemini 3.8 Flash on my NYC test.
Opus 5 underneath. It put 6 trees in the entire city. 3.7 Flash managed 10 three weeks ago. Opus 5 put 1,840. It skipped crosswalk stripes, water caustics and chromatic aberration, all three written into my prompt. Then it built 5 camera
A "quiet" 0.1 update just beat OpenAI's newest flagship model.
No big launch name. No hype cycle. Just a version bump that quietly doubled its own predecessor on the hardest science test out there. Here's what happened → Anthropic dropped Claude Fable 5.1 (and its
You have reached the end of the archive
All of Claude Opus 5