Skip to content

Anthropic

Mythos got Claude on AMD GPUs in 3 days vs 6-12 months

Written by an AI model from the linked sources and checked against them by a second model. How we work

Image: madrobot.blog

In its IPO roadshow video, Anthropic says Claude Mythos got Claude running on AMD GPUs in about three days “with basically no human intervention.” A footnote narrows that to prompt processing for one Anthropic model on AMD MI355X GPUs, and both the three days and the ”~6-12 months” baseline are Anthropic’s own figures from marketing for its share sale.

Prompt processing is the stage where a model reads its input before answering, so the footnote doesn’t cover a whole model or training. Tom Brown, co-founder and Chief Compute Officer, says Claude “ported or wrote almost 100% of the kernels,” the chip-specific programs that do a model’s math.

AMD hasn’t commented. Final pricing is expected November 18, with Nasdaq trading under ANT the next day.

Google

Gemini 4 Hydrogen reportedly reaching all Gemini app users

Per a Threads post, Gemini app users are getting Gemini 4 Hydrogen in a fast, gradual rollout to everyone, and the post’s author expects Google’s official word soon. When a new model lands in your Gemini app, do you switch right away or wait?

No comments yet. Be the first.

Join the conversation

Sign in to leave a comment. It appears after moderation.