Open Brain
VercelEngineering PostOfficial Source

Grok 4.7 now available and 40% off on AI Gateway, fx, and eve

Grok 4.7 is available to coding agents through Vercel AI Gateway with selectable reasoning depth, a 500K context window, and a temporary 40% discount.

Vercel · Sep 21, 2026
Open Source Open MarkdownOpen JSON
Source Summary

Vercel AI Gateway now offers **Grok 4.7** with a **500K-token context window** and low, medium, high, or xhigh reasoning. The same model ID works across AI SDK, Chat Completions, fx, Cursor, Codex, Amp, OpenCode, and eve.

Practical Implication

Builders can vary reasoning depth for latency-sensitive versus deeper tasks while keeping one Gateway model ID. Requests also participate in Gateway logs, reporting, budgets, routing, retries, and failover.

Agent-Ready Context
Vercel AI Gateway now offers **Grok 4.7** with a **500K-token context window** and low, medium, high, or xhigh reasoning. The same model ID works across AI SDK, Chat Completions, fx, Cursor, Codex, Amp, OpenCode, and eve.

Builders can vary reasoning depth for latency-sensitive versus deeper tasks while keeping one Gateway model ID. Requests also participate in Gateway logs, reporting, budgets, routing, retries, and failover.

The **40% discount ends September 27**. The material provides no benchmark, latency, or workload-cost results, so the promotion should not substitute for evaluation.
Connected Context · Feed7 Judgment

Grok 4.7 expands the gateway’s already crowded coding-model pool with one reusable ID and an explicit reasoning-depth control. It does not resolve selection against other long-context or tiered models: the temporary discount and stated capacity are trial inputs, while representative quality, latency, and post-promotion cost remain the deciding evidence.

Gemini 3.8 Flash now available on AI GatewayGemini 3.8 Flash is a direct gateway alternative with a larger stated context window and a different temporary discount, reinforcing that promotional pricing and capacity alone cannot choose the production route.GLM 5.3 now available on AI GatewayGLM 5.3 supplies another long-context coding route whose claimed token efficiency should be compared with Grok 4.7’s adjustable reasoning on matched repository workloads.GPT 5.6 Sol, Luna, and Terra now available on AI GatewayGPT 5.6 splits capability and cost across separate Sol, Terra, and Luna targets, contrasting with Grok 4.7’s attempt to vary reasoning depth behind one model ID.Introducing Grok 4.6Grok 4.6 is the closest supplied predecessor and provides the relevant baseline for testing whether 4.7 changes long-running coding behavior rather than assuming improvement from the newer version number.
Context Map
modelcoding#model-selection#reasoning#coding-agents
Uncertainty
The **40% discount ends September 27**. The material provides no benchmark, latency, or workload-cost results, so the promotion should not substitute for evaluation.