Latest AI news · Original source attached
xAI ships Grok 4.7 for coding and knowledge work
Grok 4.7 arrives at the same price as 4.6 with a larger base model, a longer training run, and xAI's strongest safeguards to date.
Reported from xAI (SpaceXAI)
Summary written from the original first-party source — the link to it is at the end of this post.
xAI released Grok 4.7, its most capable model for coding and knowledge work, on a new larger base model with a longer reinforcement learning run weighted toward problems that take hours to complete. It ships at the same price and speed as Grok 4.6.
What’s actually new
- Bigger model, longer training. A larger base model plus an RL run focused on multi-hour tasks, trained to natively understand the Grok Bot harness for conversational and knowledge work.
- Better coding and office work. On CursorBench 4.0, Grok 4.7 scores 46.3% at xHigh effort, up from 40.4% for Grok 4.6 at High, and it hits 71.0% on DeepSWE v1.1 at high effort.
- A new safeguard stack. The strongest refusal and jailbreak resistance xAI has tested, leading LatchBio’s biosafety benchmark at 62.4% and letting only 3.3% of risky cyber prompts through HackerBench v0.3.
Why it matters
Grok 4.7 undercuts the competition on price while staying near the frontier in coding and knowledge work. At $2 per million input tokens and $6 per million output, it is live today in Cursor, Grok Build, and the Grok API. For teams evaluating agent models, it is a serious cost efficient option in the coding class.
(Original source: xAI (SpaceXAI))
AI Slop Weekly
Get this story in context, every Sunday.
The monthly rundown of what the stories above add up to — what shipped, what changed, what to actually use. 3-minute read, free forever.
Subscribe freeNo spam. Unsubscribe anytime.