| 创客 | SpaceXSI / xAI |
|---|---|
| 已发布 | 2026 年 9 月 21 日 |
| API基础输入/输出 | 每百万代币 2 美元/6 美元 |
| 启动时访问 | Cursor、Grok Build、API、路由器和云平台 |
| 快速变体 | 以两倍的价格实现两倍的输出速度 |
发布公告。基本价格;更高上下文的费用是单独的。 · 来源
| 规格 | 公布价值 |
|---|---|
| 型号编号 | 格罗克-4.7 |
| 语境 | 500,000 个代币 |
| 输入→输出 | 文字和图像 → 文字 |
| 缓存输入 | 每百万代币 0.50 美元 |
| 推理努力 | 低、中、高、x高;默认高 |
| 长上下文定价 | 200K 以上的更高投入定价 |
文件检查日期为 2026 年 10 月 30 日;当前限制不被断言为历史启动限制。 · 来源
Unofficial, from StatusGator. Uptime is the share of the window not marked down (0 min down, 1704 min degraded); it covers the whole service, not this version alone.
Estimate from the version's published specs, benchmarks and reviews, 0-5; scores are only comparable within one criteria year. Scored as launched (Grok 4.7, 21 Sep 2026): Artificial Analysis Intelligence Index 46 at xhigh (leaders 53); Terminal-Bench 4.0 25.8% as measured by Artificial Analysis (xAI's own run: 37.6%); 500k-token context; $2 in / $6 out per 1M tokens up to 200k (about $3 blended at 3:1); 96.6 output tokens/s on Artificial Analysis; text and image input, text output. xAI published no GPQA Diamond or SWE-bench Verified figure.[1]
but still short of Fable 5.1, which wins at every price point
On Terminal-Bench 4.0, Grok 4.7 hits just 26 percent, versus 60 percent for GPT-6 Astra and 55 percent for Claude Fable 5.1
Grok 4.7 found and fixed a real, unscripted bug in a containerized full-stack app
it is the best price-performance model I have tested this quarter, but not the smartest one on the board.
On those three tasks, at default settings, we found no measurable difference in self-checking between Grok 4.7 and Grok 4.6
| xAI Launches Grok 4.7. It's Bigger, But Late to the AI Frontier Party · 2026年9月21日 | 0.30 |
|---|---|
| xAI launches Grok 4.7 at bargain prices, but benchmarks reveal a wide gap to Claude and GPT-6 · 2026年9月21日 | -0.10 |
| Grok 4.7 Hands-On: xAI's New Model Tested on Real Bugs, Circuits, Law · 2026年9月22日 | 0.50 |
| Grok 4.7 review: is xAI's cheap frontier model actually good? · 2026年9月23日 | 0.50 |
| Grok 4.7 Review: xAI's Self-Checking Model · 2026年9月25日 | 0.40 |
读者上传的照片,与图钉带来源的媒体分开显示。你可以为吃到或看到的内容评分,在餐厅图钉上还能直接在照片上标记菜品。
第一项始终是图钉的来源。总体可信度是各份资料对上方所用开始和结束时间支持程度的加权平均;资料每比最新的一份旧 180 天,权重减半。
发布公告日期为“2026 年 9 月 21 日”,文中写道:“Grok 4.7 今天已在 Cursor 和 Grok Build 中提供”;xAI 的 API 更新说明将其列在 9 月 21 日,Decrypt[4] 也称“xAI 在周一(9 月 21 日)发布了 Grok 4.7”。
The API release notes under September 21: Grok 4.7 "is now available on the xAI API as grok-4.7" with a 500k context window, text and image input, text-only output and no output limit, $2 / $0.50 / $6 per million tokens below 200k prompt tokens and $4 / $1 / $12 above, reasoning effort low to xhigh, and a 2x-priced Grok 4.7 Fast only in Cursor and Grok Build.[1]
Grok Bot Release Notes's episode "Grok 4.7 Makes Autonomous AI Teams Practical" (2026-09-23) says: "So, okay! On September 21st , 2026, SpaceXAI officially dropped Grok 4.7, and... wow, this is a massive leap for persistent background agents. The main headline here is that Grok"[1]
Decrypt's same-day report that xAI released Grok 4.7 on Monday after Musk walked the timeline back at least five times since late July; it reports 2.1 trillion parameters, up 40% from Grok 4.6's 1.5 trillion, and supplemental training on SpaceX engineering data.
The Decoder's same-day report: Grok 4.7 scores 46 on the Artificial Analysis Intelligence Index (v4.3.2) against 53 for Claude Fable 5.1 and GPT-6, and 26% on Artificial Analysis' Terminal-Bench 4.0 run against 60% for GPT-6 Astra and 55% for Claude Fable 5.1.
有遗漏或错误吗?用你自己的话说明:能佐证此图钉的链接、不同的开始或结束日期及理由,或缺失、有误的信息。AI 会对照此图钉的来源进行核实,搜索更好的来源,并添加任何支持你说法的页面。图钉自身的来源仍然最重要。AI 也会查看图片:显示的是别的东西或显示效果差的图片会被移到后面或替换。
原标题: Grok 4.7 Released
Epoch AI model database (Grok 4.7); Epoch's confidence in the figures: Likely. Blank tiles are not published.