xAI Ships Grok 4.7 at Unchanged $2/$6 Pricing, Undercutting Rivals 80% on Coding

Grok 4.7 arrives with an aggressive value proposition: matching or beating frontier coding benchmarks while keeping token prices flat at $2/$6 per million input/output tokens. xAI reports 46.3% on CursorBench 4.0 and 71% on DeepSWE v1.1, positioning the model directly against Claude Fable 5.1 and GPT-6 Astra in the coding-assistant market. The larger architecture — cited at roughly 2.1 trillion parameters in some reports — adds improved self-verification and long-context handling, and rolls out instantly to GitHub Copilot.

The pricing is the story. Multiple outlets frame Grok 4.7 as an 80% discount to comparable models, a deliberate wedge into the coding workload segment where token volume is high and margins matter. Days earlier xAI also shipped Grok Voice Transcribe 2.0 with double the accuracy of v1.0 at unchanged pricing ($0.10/hour batch, $0.20/hour streaming), reinforcing a pattern of holding prices while raising capability.

The competitive irony was not lost on the community. Forkast's headline — 'Nine days after Musk endorsed a slowdown, xAI ships a model that undercuts the slowdown by 80%' — captured the tension between the industry's public pacing rhetoric (Amodei, Altman, Musk all signaling deceleration) and the reality of a price war. Skeptics at VentureBeat flagged a real caveat: Grok 4.7's high token consumption on agentic workloads could erode the sticker-price savings, threatening real-world ROI versus Claude despite the lower per-token cost. Developers will be watching whether the effective cost, not the headline rate, actually beats incumbents in production.