September 22, 2026
Musk agrees with Grok 4.7's top-3 coding rank: independent test puts it at 26% vs 60% for GPT-6
A second independent measurement also differs from xAI: the Artificial Analysis index gives Grok 4.7 46 points versus 53 for GPT-6.
Elon Musk
@elonmusk
Exactly.
In just the last 90 days: 1. Grok 4.3 — barely top 10. “xAI is dead beyond compute leases.” 2. Grok 4.5 — massive comeback. “Maybe a chance. But Grok will never catch up to the frontier.” 3. Grok 4.6 — they’re frontier. “Still not top 3.” 4. Grok 4.7 — now top 3 in frontier coding, while being much cheaper & faster. 5. Grok 4.8 next month..... The model machine is just starting up. Grok will be the workhorse of the upcoming agentic era.
· 4M views
Five Grok releases in 90 days. The latest, 4.7, launched on 21.09.2026 and is aiming for the top-3 in coding.
A pace of one model a month disrupts the usual calculation for anyone choosing a model for an agent: while you are reading measurements for one version, the next one arrives. And for now, only xAI measures its place in the top-3.
The price has not increased. Grok 4.7 costs $2 per million input tokens, $0,5 for cached inputs, and $6 for outputs. That is what xAI charged for 4.6. For prompts longer than 200k tokens, rates double: $4, $1, and $12.
You can get the model in Cursor, in GitHub Copilot from 21.09.2026, in the API under the name grok-4.7, and for free in Grok Build at x.ai/build, where limits apply. In Copilot, it can be selected from the model list in VS Code, JetBrains, Xcode, and Copilot CLI on Pro plans and above. In Cursor, the standard window holds 256k tokens, while the long one reaches 500k. The Fast variant responds twice as fast and costs twice as much.
Measurements differ. xAI reports 46,3% on CursorBench 4.0 and 71,0% on DeepSWE v1.1. In its index, Artificial Analysis gives Grok 4.7 46 points versus 53 for Claude Fable 5.1 and GPT-6. On agentic Terminal-Bench 4.0, the gap is wider: 26% versus 60% for GPT-6 Astra and 55% for Fable 5.1.
The price and speed are immediately visible. For now, its top-3 position has to be taken at xAI's word.
xAI promises Grok 4.8 in October 2026.
Source
