Artificial Intelligence · 13.08.2026, 02:10 UTC
[AINews] SpaceXAI Grok 4.6 and Grok @Bot
| Schweregrad | info |
|---|---|
| Kategorie | Artificial Intelligence |
| Quelle | Latent Space ↗ |
| Veröffentlicht | 13.08.2026 UTC |
Sicherheitsmeldung mit Schweregrad noch nicht bewertet. Technische Details im Tab „Originaltext“; empfohlene Schritte in der Checkliste.
One of our top recurring themes of the year has been coding agents breaking containment into knowledge work, and it’s clear that the AI teammate/multiplayer/multiagent space is the next big AI battleground. With Claude Tag launching to mixed reviews and Block’s Buzz requiring a more technical user, the space was still open for a new category leader, which the now Cursor→SpaceX team has adroitly shipped to very positive reviews:This is powered by their newest model, Grok 4.6, released today as arguably the second best knowledge work model in the world (as both competitor Cognition and Elon acknowledges)… though it is surely the top by efficiency:Grok 4.6 is a confirmed 1.5T model that “builds on Grok 4.5 with a particular focus on long-running agents and more ambitious interactive and visual work”. The only training disclosure can be reproduced in full (emphasis ours):Grok 4.6 underwent a longer supplemental training run than Grok 4.5, with curated model-generated data for reasoning and advanced technical concepts, high-quality engineering data, and an improved optimizer and training recipe. This produced a stronger foundation for the SFT and RL stages that followed.We then used Grok 4.5 to regenerate the SFT trajectories across reasoning efforts, agent harnesses, and domains such as STEM, software engineering, and knowledge work, and filtered out problematic traces with model-based checks. The resulting SFT checkpoint shows strong performance and improved behavior.Grok 4.6 is trained on a wide range of agentic RL tasks, including knowledge work, general coding, and …