Glonce

xAI launches Grok 4.7 at bargain prices, but it lags on benchmarks

modelsThe Glonce editorsThe Decoder🔥58Glonce rating58Impact45Novelty40Relevance95Surprise35Players70Credibility85
xAI launches Grok 4.7 at bargain prices, but it lags on benchmarks

xAI has released Grok 4.7, described by the company as its most capable model yet for coding and knowledge work. xAI says it is built on a larger base model, trained with longer reinforcement learning, and designed to better verify its own output. Pricing is set at $2 per million input tokens and $6 per million output tokens, rates the article notes sit closer to Chinese models than to Western frontier models.

On the independent Artificial Analysis Intelligence Index (v4.3.2), which combines ten benchmarks, Grok 4.7 scores 46, landing mid-pack. Claude Fable 5.1 and GPT-6 share the lead with 53 each. The gap widens in agentic coding: on Terminal-Bench 4.0, Grok 4.7 reaches 26 percent, well behind GPT-6 Astra at 60 percent and Claude Fable 5.1 at 55 percent. The cheaper DeepSeek V4.1 Flash also edges past Grok 4.7 on that benchmark, scoring 27 percent. Grok 4.7 is available through the Grok API, Cursor, and Grok Build.

Key facts

  • Grok 4.7 is priced at $2 per million input tokens and $6 per million output tokens, rates the article says are closer to Chinese models than Western frontier models.
  • On the Artificial Analysis Intelligence Index v4.3.2 (ten benchmarks combined), Grok 4.7 scores 46, versus a joint lead of 53 for Claude Fable 5.1 and GPT-6.
  • On Terminal-Bench 4.0, Grok 4.7 scores 26 percent, versus 60 percent for GPT-6 Astra and 55 percent for Claude Fable 5.1.
  • The cheaper DeepSeek V4.1 Flash outscores Grok 4.7 on Terminal-Bench 4.0, at 27 percent.
  • Grok 4.7 is available through the Grok API, Cursor, and Grok Build.

Why it matters

Grok 4.7 arrives priced well under the Western frontier norm, more in line with Chinese model pricing, while xAI pitches it as its strongest coding and knowledge model so far. That combination puts pressure on the assumption that top-tier capability requires top-tier pricing, even though the benchmarks below show Grok 4.7 is not matching the capability of the models it is priced against.

Who it affects

Developers and teams choosing an API for coding or knowledge work, particularly those weighing cost against raw capability. Grok 4.7 is already reachable through the Grok API, Cursor, and Grok Build, so anyone building on those tools can access it directly rather than waiting on a separate integration.

How to use it

Grok 4.7 is available now through the Grok API at $2 per million input tokens and $6 per million output tokens, and through Cursor and Grok Build. No separate release date or rollout schedule is stated beyond its current availability.

How solid is it

The benchmark figures come from two named, independent sources: the Artificial Analysis Intelligence Index (v4.3.2), which combines ten benchmarks, and Terminal-Bench 4.0 for agentic coding. Both give Grok 4.7 a clear, specific ranking against Claude Fable 5.1, GPT-6, and DeepSeek V4.1 Flash rather than a vague comparison, which makes the mid-pack and agentic-coding results concrete rather than impressionistic.

Risks and caveats

xAI's own claims about the model, that it uses a larger base model, longer reinforcement learning, and better output verification, are attributed to the company and are not independently verified in the source. The article's explanation for the low pricing ("probably for good reason") is its own inference, not a stated rationale from xAI. No release date and no author byline appear in the source, and no technical detail is given on what the longer reinforcement learning or output verification actually involve.