Grok 4.7

xAI has introduced Grok 4.7 on its official news page (dated Sep 21, 2026), describing it as its most capable model for coding and knowledge work, built on a new larger base model than Grok 4.6, trained with a longer reinforcement learning run on harder, multi-hour tasks. According to the announcement, Grok 4.7 is bett Pricing is set at $2 per million input tokens and $6 per million output tokens, the same as Grok 4.6, with the announcement claiming it is "twice as fast, at half the price of comparable models." Published benchmark scores include 46.3% on CursorBench 4.0, 71.0% on DeepSWE v1.1, 64.0% on EEBench, 38.0% on Terminal-Benc

According to the DataCamp write-up, SpaceXAI released Grok 4.7 on September 21, 2026 as the successor to Grok 4.6, built on a larger base model with a longer reinforcement-learning run weighted toward multi-hour tasks. Pricing is reported at $2 per million input tokens and $6 per million output tokens, unchanged from G The article frames Grok 4.7 as a value-oriented release, noting that Fable 5.1 still leads on absolute coding and clinical reasoning at $10/$50 per million tokens, making Grok 4.7 an eighth of that output price. It states Grok 4.7 was trained to work longer on hard tasks and to verify its own work, with the generationa

The llm-stats.com model page tracks Grok 4.7's benchmark scorecard sourced from x.ai, ranking it 22nd on the composite LLM Stats Score. Per the excerpt, it ranks 13th of 270 on Coding (top half), 41st of 196 on Tool Calling, 68th of 365 on Reasoning, and 101st of 245 on Healthcare. Reported scores include GDPval Elo 16 The page situates Grok 4.7 against peers on a price-versus-score plot, with comparable points including Gemma 4 E4B at $0.024 (score 13.8), GPT OSS 120B at $0.043 (28.8), DeepSeek-V4-Flash-0731 at $0.066 (44.5), DeepSeek-V4.1-Flash at $0.24 (51.8), Grok 4.7 at $2.19 (48.5), and GPT-6 Astra at $11.9 (59.7). The data was

The Decoder reports (Sep 21, 2026) that xAI launched Grok 4.7, its most capable model yet for coding and knowledge work, built on a larger base model with longer reinforcement learning training and designed to better verify its own output. Pricing is set at $2 per million input tokens and $6 per million output tokens, On the independent Artificial Analysis Intelligence Index (v4.3.2), which combines ten benchmarks, Grok 4.7 scores 46 and lands mid-pack, behind Claude Fable 5.1 and GPT-6 which both lead with 53. On Terminal-Bench 4.0's agentic coding test, Grok 4.7 hits just 26 percent, compared to 60 percent for GPT-6 Astra and 55 p

A Cursor Community Support Engineer posted a Main Announcement on September 21, 2026 confirming that Grok 4.7 is now available through Cursor's Composer integration and soliciting developer feedback on long-running tasks, non-code knowledge work, and visual/interactive projects compared with Grok 4.6. The post names Gr The Cursor thread highlights that Grok 4.7 was trained to natively understand the agent harness, suggesting it should be noticeably better at conversational tasks, research, and producing documents and presentations, and invites users to compare Grok 4.7 against Grok 4.6, Composer, and other frontier models in their da

Unite.AI reports (dated Sep 21, 2026) that SpaceXAI released Grok 4.7, its newest model for coding and knowledge work, priced from $2 per million input tokens and $6 per million output tokens. The article states Grok 4.7 uses a new larger base model than Grok 4.6, was trained with a longer reinforcement learning run we The article cites benchmark numbers including 46.3% on CursorBench 4.0 (vs. 40.4% for Grok 4.6, 41.7% for GPT-5.6 Sol, and 51.8% for Fable 5.1) and notes the company claims Grok 4.7 is twice as fast at half the price of comparable models. It also references Grok 4.6's prior extensions to GitHub Copilot, Amazon Bedrock,

Artificial Analysis published a dedicated benchmarking analysis of Grok 4.7 on September 21, 2026, evaluating it at xhigh reasoning effort. Grok 4.7 scores 46 on the Artificial Analysis Intelligence Index, a +2-point gain over Grok 4.6, bringing SpaceXAI into the top 4 AI labs on the index and overtaking GPT-5.6 Sol on On the Coding Agent Index, Grok 4.7 paired with Grok Build scores 56 (+9 vs. Grok 4.6 xhigh), ranking 4th among native harnesses behind Claude Fable 5.1, GPT-6 Astra, and Claude Opus 5. Terminal-Bench 4.0 improved +4.5 pp and GDP.pdf +3.0 pp, with small regressions on AA-LCR (-3.7 pp) and AutomationBench-AA (-1.1 pp).

CellCog confirms Grok 4.7 shipped on September 21, 2026, forty days after Grok 4.6, citing xAI's launch post on x.ai/news, the model docs listing grok-4.7, and a @SpaceXAI announcement at 16:17 UTC. The model is available in Cursor, Grok Build, and the Grok API, plus third-party harnesses, routers, and cloud platforms. On xAI's own benchmark table, Grok 4.7 beats Grok 4.6 on all seven listed benchmarks and tops GPT-5.6 Sol Max on four; Claude Fable 5.1 Max leads it on CursorBench, Terminal-Bench, GDPval, and HealthBench, while Grok 4.7 leads Fable on EEBench and the Harvey legal benchmark. Artificial Analysis's first independent read

Artificial Analysis's model profile page for Grok 4.7 (xhigh) lists it as a proprietary SpaceXAI model released in September 2026, with a 500k-token context window, text-and-image input, text output, and configurable reasoning effort. It scores 46 on the Intelligence Index (well above the 26 median), with a 44.2 tokens Pricing is set at $2.00 per 1M input tokens and $6.00 per 1M output tokens (both described as moderately priced, at or below medians of $2.00 and $10.00), with cache hits at $0.50 per 1M, and a per-task Intelligence Index cost of $3.74. The model is positioned in the 4-out-of-4 tier for both intelligence and cost among

SpaceXAI's own news page lists "Introducing Grok 4.7" dated September 21, 2026, describing it as the company's most powerful model for coding and knowledge work, with the marketing claim of being twice as fast at half the price of comparable models. The page excerpt serves as the first-party confirmation of Grok 4.7's The same index also documents surrounding Grok product activity in August and September 2026 — including Grok Bot availability across SuperGrok, Cursor Pro, and Cursor Teams plans, Grok 4.6 placements on Microsoft Foundry, Gemini Enterprise Agent Platform, Amazon Bedrock, and GitHub Copilot, and Grok Voice Transcribe 2