Grok 4.7 Coding Agent Scores Jump to 56, Led by Terminal-Bench Surge
As seen on the 24/7 Wall St. homepage on September 21, 2026.
xAI's Terminal-Bench score nearly doubling in one model generation is the kind of jump that decides which lab enterprises hand their coding budgets to.
Grok Build with Grok 4.7 (xhigh) scores 56 on the Artificial Analysis Coding Agent Index, up from 47 with Grok 4.6 (xhigh). It improves across all three components: DeepSWE v1.1 rises from 65% to 73%, Terminal-Bench 4.0 from 18% to 33%, and SWE-Atlas-QnA from 58% to 63%. These https://t.co/JYfH3Uyi82
- Replies1
- Reposts1
- Likes22
Continue ReadingShow less
Grok Build's overall score on the Artificial Analysis Coding Agent Index climbed from 47 to 56 when moving from Grok 4.6 to Grok 4.7, both tested at the xhigh compute tier. That nine-point gain is broad-based, with every component of the index moving higher.
The headline mover within the index is Terminal-Bench 4.0, where the score jumped from 18% to 33%. That near-doubling in a single model generation is unusually large for a benchmark designed to stress-test autonomous, command-line-level coding tasks, the kind that enterprise software teams rely on most heavily for real workflow automation.
Sponsored
Twelve Tabs, One Thesis
Your Research Resets Every Morning
The quote page in one tab. Filings in another. A chart you rebuilt from scratch, a transcript you never went back and found, a screener whose settings you will redo next week. Nothing you built yesterday is still there.
AlphaSpace replaces all of it with one screen you arrange yourself. Earnings calendar, estimate versus actual, the call transcript, live news, your own charts, every panel wired to whatever ticker you click. Close the browser and it is all still sitting there tomorrow.
See What a Built View Looks Like →
(Sponsor)
The other two components also improved in a meaningful way, with DeepSWE v1.1 and SWE-Atlas-QnA both moving higher. Taken together, the gains point to a genuine lift in underlying capability across task types.
For enterprises evaluating which AI coding platform to budget for, a composite index score that reflects terminal, software engineering, and question-answering tasks carries more weight than any single-task leaderboard. The degree to which Grok 4.7 closes or extends the gap versus competing labs on this index will be a practical factor in procurement decisions going forward.