Grok 4.6 ships in Cursor and Grok Build

xAI has released Grok 4.6, a model the lab says has pulled even with GPT-5.6 Sol on the composite Artificial Analysis Intelligence Index. According to the announcement posted on Threads, Grok 4.6 also runs level with Fable 5 and GPT-5.6 Sol across many benchmarks.

At a glance

• Grok 4.6 builds on Grok 4.5, with xAI focusing the update on long-running agents and on more ambitious interactive and visual work, the two areas the lab names for this version.
• Frontend Arena lists the model in 13th place with 1553 points, a ranked result that sits next to the composite index score xAI cites for the parity claim.
• Access starts in Cursor and Grok Build on day one, with double usage included for the first week, which puts the model into everyday coding work before independent testing of the claims.

Why it matters Parity claims against a rival flagship are easy to state and slow to verify, and a composite index can average away wide gaps on individual tasks. The Frontend Arena placement reads as the harder data point of the two, since it is a ranked position rather than a self-reported comparison. Shipping directly into Cursor likely matters more than either figure, because that is where the claim meets daily use.

Grok 4.6 is available in Cursor and in Grok Build from the day of the announcement. Both include double the usual usage for the first week, an allowance that lapses after that window. No pricing details or standard rate-limit figures accompanied the release.

On the benchmark side, xAI puts Grok 4.6 at the level of GPT-5.6 Sol on the Artificial Analysis Intelligence Index, a composite score rather than a single test. The lab describes the model as running level with Fable 5 and GPT-5.6 Sol across many benchmarks, without naming the individual tests in the announcement.

The ranked placement published with the release is on Frontend Arena, where Grok 4.6 sits in 13th with 1553 points. The announcement gives the score without the surrounding leaderboard, so the gap to the models above it is not stated.

Grok 4.6 builds on Grok 4.5, which xAI had already opened to every user of X and its apps. xAI says this version is aimed at long-running agents and at more ambitious interactive and visual work, the two capability areas named in the release.

What's next The near-term test is whether outside runs of the Artificial Analysis index reproduce the parity with GPT-5.6 Sol that xAI reports. Whether the doubled usage in Cursor and Grok Build continues past the first week is not addressed in the announcement, and no pricing or wider rollout schedule for Grok 4.6 has been given.