xAI launched Grok 4.7 on September 21, 2026, describing it as its most capable model for coding and knowledge work. The model is built on a new, larger base architecture compared to Grok 4.6, trained with a longer reinforcement-learning run weighted toward multi-hour tasks, and is designed to natively understand the Grok Bot harness for conversational and general knowledge work.
Grok 4.7 is generally available as of the announcement date. xAI states it is accessible in Cursor, Grok Build, the Grok API, third-party coding harnesses, and model routers and cloud platforms. Base pricing starts at $2 per million input tokens and $6 per million output tokens, with a fast variant offered at twice the output speed and twice the price. xAI notes the base tier is served at the same price and speed as Grok 4.6.
In vendor-reported benchmark results, xAI positions Grok 4.7 as competitive in its class. On CursorBench 4.0, xAI reports a score of 46.3% for Grok 4.7 (xHigh setting) versus 40.4% for Grok 4.6, 41.7% for GPT-5.6 Sol, and 51.8% for Fable 5.1 Max. On DeepSWE v1.1, xAI reports 71.0% (high-effort) for Grok 4.7. On EEBench, the reported score is 64.0%. For multi-hour office work on AA Briefcase v1.1, xAI reports a score of 1,657. On Terminal-Bench 4.0, the reported figure is 37.6%. On the Harvey Legal Agent Benchmark, xAI reports 19.6%, and on HealthBench Professional, 56.7%. xAI characterizes the model as being "at the frontier in price-performance" on CursorBench 4.0.
On safety, xAI states Grok 4.7 uses an entirely new safeguard stack and is its strongest model tested on refusals and jailbreak resistance. xAI reports a 62.4% score on LatchBio's biosafety benchmark and states the model allows only 3.3% of risky dual-use prompts through on HackerBench v0.3, which xAI describes as its own benchmark for risky and malicious cyber tasks. xAI also announced invite-only access to Grok 4.7's red-team capabilities for select cybersecurity partners conducting defense research.
All performance figures, pricing comparisons, and safety metrics above are as reported by xAI in its launch announcement. No independent third-party verification of the benchmark scores or safety claims is included in the supplied evidence.