Amazon Bedrock adds Grok 4.7 with adjustable reasoning effort

AWS announced that SpaceXAI's Grok 4.7 is available on Amazon Bedrock, with tool calling, structured output and four configurable reasoning effort levels. SpaceXAI lists a 500K-token context window, text and image input, and text output. AWS tested an agent on three synthetic tasks involving outage diagnosis, maintenance scheduling and migration selection. It answered 67 of 72 runs correctly. The highest effort level answered all 18 runs correctly, compared with 15 of 18 at low effort; output tokens rose from 1,184 at low to 9,147 at the highest level. Access uses cross-Region inference profiles. AWS cautioned that six runs per task and effort level make the evaluation indicative rather than statistically conclusive.