Grok 4.7.
Grok 4.7 is SpaceXAI's flagship model for coding, agentic tasks, and knowledge work, succeeding Grok 4.6. It is particularly strong at long-running software engineering tasks, verifying its own work, and...
No self-host path — closed weights.
Grok 4.7's weights aren't published. Use it via the access providers below.
Cheapest hosted endpoints.
| Provider | Access | $/M in | $/M out | |
|---|---|---|---|---|
| OpenRouter | api aggregator | $2.0 | $6.0 | Launch ↗ |
Variants in the Grok family.
xAI's frontier model with built-in DeepSearch + real-time X integration.
Grok 4.6 is SpaceXAI's smartest model with frontier performance on coding, kn...
Grok 4.5 is SpaceXAI's smartest model with frontier performance on coding, kn...
Grok 4.20 is an xAI reasoning model with text and image input and a 1M-token ...
Grok 4.20 Multi-Agent is a beta research model that runs 4 or 16 agents behin...
Grok 4.3 is an xAI API model with text and image input and a 1M-token context...
What it costs per month across providers.
Estimate your monthly bill for Grok 4.7 across every host that publishes per-token pricing. Slide your token volumes; the chart + table re-rank cheapest-first.
最便宜的服务商在左侧。
每月总成本 — 输入与输出 token 合计。
账单明细。
| 服务商 | 月度总计 | |
|---|---|---|
| $32.0 | 注册 ↗ |