Come join us to work on 🔢🧮 for 🔄! x.com/i/jobs/1968405…

We are hiring numerics / quantization expert to scale RL (with sehoonkim)! there will be lots of exciting challenges coming with Jax + Sglang + the first Gigawatt cluster in the world (with many hundred thousands of GB200/300)!

Looking for 🔢quantization / numerics experts 🔢 for scaled RL training! Quantization is no longer a good-to-have feature, but the key component for building strong intelligence with our world-best RL infra + gigawatt cluster🔥

Introducing Grok 4 Fast, a multimodal reasoning model with a 2M context window that sets a new standard for cost-efficient intelligence. Available for free on grok.com, grok.x.com, iOS and Android apps, and OpenRouter. x.ai/news/grok-4-fa…

Grok 4 Fast is out, for free for everyone! It pushes the frontier of reasoning efficiency and search 🚀 Join our team to maximize intelligence density x.com/i/jobs/1948829…

The RL Infrastructure team xAI rolled out a new agent framework internally which was at core of the training run for Grok 4 Fast. It will also be at the core of all our future RL training runs and we need the best talent in the world to fully harness the power of Colossus 2 👀.

At xAI, we are building the world’s most advanced inference system on tens of thousands of GPUs. It has been a fun journey to support the Grok 4 Fast long-context model end-to-end, from autoscaling, disaggregated serving, to model parallelism. Please DM me or apply to the below

Deterministic inference, here you are. True on-policy RL is on the way. Although we are mostly using off-policy, having a deterministic mode will make many things easier!

The main branch of sglang now supports deterministic inference with user-specified per-request seeds! It utilized kernels from Thinking Machines and introduced new optimizations & coverage. Run out of box for most hardware backends and pytorch versions.

Hiring for a new team building computer control agents. Join us to build Grok5 / macrohard later this year. DM me! Will send out a job post soon too.

We are hiring to build an AI-native knowledge base/search engine. Apply here if interested: job-boards.greenhouse.io/xai/jobs/49142…

Today, Ekin Dogus Cubuk and I are excited to introduce Periodic Labs. Our goal is to create an AI scientist. Science works by conjecturing how the world might be, running experiments, and learning from the results. Intelligence is necessary, but not sufficient. New knowledge is

iteration speed is a superpower

I have been waiting for this update! Finally, you can unlock the full potential of Grok 4 and Grok 4 Fast. So much of our training compute involves tool use and agents–––now you can harness Grok exactly as it's built. 🔎🔧🦾

Grok 4 Fast has the best intelligence density 🚀

Really proud to see Ray have become the Linux foundation project! It’s been almost 6 years since I first started working on Ray. Still remember when I first worked on the project 100% of tests are flaky lol…

proud of both the companion and imagine team. We didn't have any usable video gen models 3 months ago and most efforts started 3 months ago.

🚀