# Grok 4.7

Grok 4.7 xHigh is a proprietary large language model by xAI, released in June 2026, ranked 15th on LMArena web development leaderboards and 22nd on LiveBench overall.

Grok 4.7 xHigh is a proprietary [large language model](https://www.wikiprompt.org/wiki/large-language-model) developed by [xAI](https://www.wikiprompt.org/wiki/grok), released on 2026-06-25. It is a variant of the Grok 4.7 series, positioned as a high-compute configuration for demanding tasks. The model is proprietary and has been evaluated on public benchmark leaderboards, including [LMArena](https://www.wikiprompt.org/wiki/lmarena) and LiveBench, where it has achieved notable rankings as of the 2026-10-10 snapshot.

Grok 4.7 xHigh is part of the broader [Grok 4](https://www.wikiprompt.org/wiki/grok-4-6) lineage, succeeding earlier versions such as [Grok 4.5](https://www.wikiprompt.org/wiki/grok-4-5) and [Grok 4.6](https://www.wikiprompt.org/wiki/grok-4-6). It is designed for advanced reasoning, coding, and agentic workflows, building on the architectural innovations of its predecessors. The model is available through xAI's proprietary platforms, with no open-source release.

## Benchmark Performance

On the LMArena web development leaderboard, Grok 4.7 xHigh ranks 15th overall with a score of 1639.09, and 15th in the webdev category with the same score. In the webdev-react subcategory, it achieves a higher score of 1661.15, also ranking 15th. For webdev-html, it ranks 19th with a score of 1625.24. In the image-to-webdev task, the model ranks 23rd with a score of 1577.34.

On the LMArena agent leaderboard, the variant is listed as "Grok 4.7 (xHigh)" and ranks 21st overall, though no score is reported. On LiveBench, Grok 4.7 xHigh ranks 22nd overall with a score of 78.34. In the Artificial Analysis intelligence index, the xHigh variant ranks 29th with a score of 46.4, while the High variant ranks 30th with a score of 46.3.

## Architecture and Training

Grok 4.7 xHigh is built on a [transformer](https://www.wikiprompt.org/wiki/transformer) architecture, consistent with modern [deep learning](https://www.wikiprompt.org/wiki/deep-learning) approaches. It likely employs techniques such as [residual connections](https://www.wikiprompt.org/wiki/residual-network), [Adam optimization](https://www.wikiprompt.org/wiki/adam-optimizer), and [gradient clipping](https://www.wikiprompt.org/wiki/gradient-clipping) to stabilize training. The model is trained on large-scale datasets, though specific details on data composition and compute are not publicly disclosed.

The "xHigh" designation suggests a higher compute allocation or larger parameter count compared to standard Grok 4.7 variants, potentially enabling improved performance on complex tasks like code generation and web development. The model is optimized for agentic use cases, as evidenced by its presence on the LMArena agent leaderboard.

## Development and Release

Grok 4.7 xHigh was released on 2026-06-25, following the release of [Grok 4.6](https://www.wikiprompt.org/wiki/grok-4-6) and other intermediate versions. It is developed by xAI, a company founded by [Elon Musk](https://www.wikiprompt.org/wiki/grok) in 2023, which has focused on creating AI systems that are "maximally curious" and capable of advanced reasoning. The model is part of xAI's strategy to compete with other frontier labs like [OpenAI](https://www.wikiprompt.org/wiki/openai) and [Anthropic](https://www.wikiprompt.org/wiki/anthropic).

The release of Grok 4.7 xHigh coincided with updates to xAI's infrastructure, including integration with [Grok Bot](https://www.wikiprompt.org/wiki/grok-bot) and other products. The model is available through xAI's API and consumer-facing platforms, though access may be limited to certain tiers.

## Comparison with Other Models

Grok 4.7 xHigh competes with other high-performance models such as [GPT-5.5 xHigh Codex Harness](https://www.wikiprompt.org/wiki/gpt-5-5-xhigh-codex-harness) and [GPT-5.6 Luna xHigh Codex Harness](https://www.wikiprompt.org/wiki/gpt-5-6-luna-xhigh-codex-harness). On LMArena webdev, it ranks 15th, while these GPT variants are not listed in the provided data, suggesting they may not be directly comparable or are ranked differently. On LiveBench, Grok 4.7 xHigh's score of 78.34 places it 22nd, indicating strong but not top-tier performance.

Compared to earlier Grok models, such as [Grok 4.1](https://www.wikiprompt.org/wiki/grok-4-1) and [Grok 4.1 Fast](https://www.wikiprompt.org/wiki/grok-4-1-fast), the 4.7 xHigh variant shows improved scores, reflecting ongoing advancements in xAI's training methods. The model's performance on web development tasks, particularly React, suggests a focus on practical coding applications.

## Reception and Impact

Grok 4.7 xHigh has been noted for its competitive performance on public benchmarks, though it does not lead any major leaderboard. Its ranking on LMArena and LiveBench indicates it is a capable model, but it faces stiff competition from models like [GPT-5.6 Sol](https://www.wikiprompt.org/wiki/gpt-5-6-sol-xhigh-codex-harness) and [GPT-5.6 Terra](https://www.wikiprompt.org/wiki/gpt-5-6-terra-xhigh-codex-harness). The model's proprietary nature limits independent analysis, but its presence on multiple leaderboards provides some transparency.

The model is part of a broader trend in [generative AI](https://www.wikiprompt.org/wiki/generative-ai) where companies release multiple variants to cater to different performance and cost needs. Grok 4.7 xHigh's high-compute configuration likely targets enterprise users who require maximum capability, while lower variants may be more accessible.

## Future Directions

As of the 2026-10-10 snapshot, Grok 4.7 xHigh remains a current model, with no announced successor. xAI continues to iterate on its Grok series, and future versions may incorporate feedback from benchmark evaluations. The model's performance on agentic tasks suggests that xAI is investing in [reinforcement learning from AI feedback](https://www.wikiprompt.org/wiki/rlaif) and other techniques to improve reasoning and tool use.

Given the rapid pace of AI development, it is likely that Grok 4.7 xHigh will be superseded by newer models within months, but its benchmark results provide a baseline for measuring progress in the field.

---
Source: https://www.wikiprompt.org/wiki/grok-4-7
License: CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/)
Last updated: 2026-10-10T20:32:39.965829+00:00