started · updated
SpaceXAI launches Grok 4. 7 AI model for coding and professional tasks
SpaceXAI has officially launched Grok 4. 7, its most advanced large language model to date, specifically optimized for coding, knowledge work, and complex professional tasks. The model features a larger base architecture than its predecessor, Grok 4. 6, and utilizes extended reinforcement learning to handle tasks that may take several hours to complete. It includes a 500,000-token context window and improved self-verification capabilities.
In terms of performance, Grok 4. 7 achieved a 46.3% score on the CursorBench 4. 0 benchmark, surpassing GPT-5. 6 Sol. It also demonstrated high proficiency in electrical engineering (64.0% on EEBench) and legal tasks. While it trails some competitors in general intelligence indices and specific terminal-based coding benchmarks, SpaceXAI emphasizes its superior price-to-performance ratio. The API is priced at $2 per million input tokens and $6 per million output tokens, which is significantly lower than comparable frontier models from Anthropic and OpenAI.
Additionally, the model is natively integrated with Grok Bot, an AI agent capable of executing tasks within user applications like Gmail or LinkedIn. SpaceXAI is also utilizing Grok Bot internally to manage customer support, reporting that the agent can resolve tickets for as little as $0.20 to $0.30. The release of Grok 4. 7 follows a period of rapid development, with Elon Musk noting that the company has reached the top tier of agentic programming capabilities.
Entities
Anthropic · Elon Musk · Grok 4. 7 · Grok 4.7 · OpenAI · SpaceX · SpaceXAI · Tesla
Claims
What the coverage asserts, and how many sources carry each claim.
- [● 19 SOURCES] The API pricing for Grok 4. 7 is $2 per million input tokens and $6 per million output tokens. ihal.it · www.sbctv.gr · www.absolutegeeks.com · forkast.news · www.itmedia.co.jp · +14 more
- [● 7 SOURCES] SpaceXAI claims Grok 4. 7 is twice as fast and half the price of comparable frontier models. www.itmedia.co.jp · news.az · www.ithome.com · bitnewsbot.com · www.europesays.com · +2 more
- [● 21 SOURCES] SpaceXAI released Grok 4. 7, a model optimized for coding, knowledge work, and long-running professional tasks. bitnewsbot.com · ihal.it · www.prodiris.fr · www.sbctv.gr · www.absolutegeeks.com · +16 more
- [● 11 SOURCES] Grok 4. 7 is available through Cursor, Grok Build, the Grok API, and various third-party platforms. ihal.it · www.itmedia.co.jp · tbreak.com · www.ithome.com · news.az · +6 more
- [● 10 SOURCES] Grok 4. 7 achieved a 46.3% score on the CursorBench 4. 0 benchmark. bitnewsbot.com · www.absolutegeeks.com · www.itmedia.co.jp · pulse2.com · zdnet.co.kr · +5 more
- [● 13 SOURCES] Grok 4. 7 is built on a larger base model than its predecessor, Grok 4. 6. ihal.it · www.prodiris.fr · www.itmedia.co.jp · news.az · pulse2.com · +8 more
- [● 5 SOURCES] On the HackerBench v0. 3 evaluation, Grok 4. 7 allowed 3.3% of risky dual-use prompts through. www.sbctv.gr · www.absolutegeeks.com · www.itmedia.co.jp · www.ithome.com · ai.cnmo.com
- [● 4 SOURCES] The model scored 64.0% on the EEBench for electrical engineering tasks. www.itmedia.co.jp · pulse2.com · zdnet.co.kr · www.watch.impress.co.jp