xAI Launches Grok 4.7 for Coding and Knowledge Work
Original: Grok 4.7
Why This Matters
Grok 4.7 pushes frontier-level coding performance at a fraction of competitor pricing, intensifying cost pressure across the LLM market.
xAI released Grok 4.7 on September 21, 2026, positioning it as its most capable model for coding and professional knowledge work. Priced at $2 per million input tokens and $6 per million output tokens — same as Grok 4.6 — it scores 46.3% on CursorBench 4.0 and leads rivals on EEBench and the Harvey Legal Agent Benchmark.
xAI's Grok 4.7 is built on a larger base model than Grok 4.6, trained with an extended reinforcement learning run weighted toward tasks requiring many hours to complete. The company says the model handles longer context better, verifies its own outputs more carefully, and natively understands the Grok Bot harness for conversational and general knowledge tasks.
On CursorBench 4.0, Grok 4.7 scores 46.3% — above GPT-5.6 Sol's 41.7% but below Fable 5.1's 51.8%. It leads the field on EEBench (64.0%) and the Harvey Legal Agent Benchmark (19.6%, versus GPT-5.6 Sol's 2.5%). Pricing holds at $2/$6 per million input/output tokens, matching Grok 4.6 — considerably cheaper than GPT-5.6 Sol ($4/$20) and Fable 5.1 ($10/$50).
On safety, xAI says Grok 4.7 uses a fully rebuilt safeguard stack. It tops LatchBio's biosafety benchmark at 62.4% and on HackerBench v0.3 allows only 3.3% of risky cyber prompts through while keeping false-positive blocks low. Select cybersecurity partners are getting invite-only access to red-team capabilities for defense research.
Grok 4.7 is available now in Cursor, Grok Build, and through the Grok API.