xAI launches Grok 4.7, claiming twice the speed at half the price of comparable models
2026-09-21
xAI released Grok 4.7 on Sept 21, 2026, marketed on its news page as its 'most powerful model for coding and knowledge work,' with xAI claiming it is twice as fast and half the price of comparable models. No independent benchmark data accompanied the release.
Significance 4: A flagship price/speed claim from a tracked lab that could move cost and speed axes relevant to the bulk-cheap and coding-agent picks, but it is lab-reported with no independent verification, capping it at 4 per the rubric.
Implications · machine-drafted, not owner judgment
If xAI's price/speed claims hold up under independent testing, Grok 4.7 becomes a credible new option for the bulk-cheap and coding-agent categories, intensifying the price competition among frontier labs already visible around DeepSeek V4 and GPT-5.2. It also strengthens xAI's pitch to the developer-tool integrations it already ships in (GitHub Copilot, Cursor, cloud marketplaces) at a moment when coding-agent economics are under scrutiny elsewhere in today's material.
- An independent pricing/speed benchmark (e.g., Artificial Analysis) confirming or refuting the 'twice as fast, half the price' claim
- Grok 4.7 appearing in a tracked verdict's evidence within the next month
Sources
- primary xAI — News: Introducing Grok 4.7 retrieved 2026-09-22
Affected axes
Cost · Output speed · Coding · Hard reasoning