---
format: "aidr-story-markdown/v1"
id: "368a2b8302b90dd972e20df2e5912eeeb1179050f4b6af8c55d23162b3fedc83"
canonical_url: "https://aidr.today/368a2b83?lang=en"
title: "Grok 4.7 Puts SpaceXAI in Top 4 AI Labs With 56 Coding Agent Index Score"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-09-21T19:02:38.000Z"
category: "Models"
topics: ["anthropic","claude","multi-agent","open-source","grok","spacexai","coding-agents","benchmark"]
source_urls: ["https://huggingnews.com/ai/grok-47-puts-spacexai-in-top-4-ai-labs-with-56-coding-agent-index-score-ca998e18","https://x.com/SpaceXAI/status/2102069815225586149","https://x.com/mark_k/status/2102071475414986981","https://x.com/scaling01/status/2102067894863409493","https://x.com/ArtificialAnlys/status/2102074898327932987","https://x.com/Muennighoff/status/2102079643310571829","https://x.com/ValsAI/status/2102086608476590432","https://x.com/ValsAI/status/2102086611043508332"]
summary: "SpaceXAI released the Grok 4.7 AI model with pricing starting at $2 per million input tokens and $6 per million output tokens. The new version ranks fourth in the Artificial Analysis Coding Agent Index with a score of 56, overtaking GPT 5.6 Sol in coding agent performance. The model is available via API, Cursor, and Grok Build, featuring a 500k token context window and a rebuilt safeguard system designed for cybersecurity work. Technical improvements include a larger base model of approximately 2 trillion parameters trained with reinforcement learning for multi hour tasks. While the model scored 46 on the Intelligence Index, it dropped to 24th place on the Vals Index with a 54.2% score, compared to the 14th place ranking of Grok 4.6. These gains correlate with higher token consumption, with the model using 81k output tokens per Intelligence Index task, up from 36k for the previous version."
---

# Grok 4\.7 Puts SpaceXAI in Top 4 AI Labs With 56 Coding Agent Index Score

> [Open the canonical story](<https://aidr.today/368a2b83?lang=en>)

**Published:** 2026-09-21T19:02:38.000Z
**Category:** Models
**Topics:** anthropic, claude, multi\-agent, open\-source, grok, spacexai, coding\-agents, benchmark

## Summary

SpaceXAI released the Grok 4\.7 AI model with pricing starting at $2 per million input tokens and $6 per million output tokens\. The new version ranks fourth in the Artificial Analysis Coding Agent Index with a score of 56, overtaking GPT 5\.6 Sol in coding agent performance\. The model is available via API, Cursor, and Grok Build, featuring a 500k token context window and a rebuilt safeguard system designed for cybersecurity work\. Technical improvements include a larger base model of approximately 2 trillion parameters trained with reinforcement learning for multi hour tasks\. While the model scored 46 on the Intelligence Index, it dropped to 24th place on the Vals Index with a 54\.2% score, compared to the 14th place ranking of Grok 4\.6\. These gains correlate with higher token consumption, with the model using 81k output tokens per Intelligence Index task, up from 36k for the previous version\.

## Sources

- [Story source](<https://huggingnews.com/ai/grok-47-puts-spacexai-in-top-4-ai-labs-with-56-coding-agent-index-score-ca998e18>)
- [Story source](<https://x.com/SpaceXAI/status/2102069815225586149>)
- [Supporting source](<https://x.com/mark_k/status/2102071475414986981>)
- [Supporting source](<https://x.com/scaling01/status/2102067894863409493>)
- [Supporting source](<https://x.com/ArtificialAnlys/status/2102074898327932987>)
- [Supporting source](<https://x.com/Muennighoff/status/2102079643310571829>)
- [Supporting source](<https://x.com/ValsAI/status/2102086608476590432>)
- [Supporting source](<https://x.com/ValsAI/status/2102086611043508332>)