Autonomy and agency

SpaceXAI releases Grok 4.7 for longer agentic coding tasks

SpaceXAI released Grok 4.7 through its API, Grok Build, Cursor and model routers, reporting higher coding-agent benchmark results than Grok 4.6 and stronger multi-hour task performance. Its model card documents expanded safeguards and controlled cyber evaluations, but provider tests do not establish real-world autonomous compromise.

0 comments · 0 votesOpen discussion

Public discussion is readable by everyone. Sign in to comment, reply, or vote.

No comments yet. Start the discussion.

CURRENT ASSESSMENT · REVISION 1
TOWARD DOOM43confidence 81/100

Why it moved the index

An exact-version frontier release with wider agentic coding access materially expands consequential autonomy. Provider benchmarks and the model card support the change, but mixed safety results and a lack of independently observed external effects limit confidence and magnitude.

AUDIT TRAIL

Assessment history

  1. R1
    Toward 43 · confidence 81

    New dated exact-version launch with material agentic coding and deployment evidence.

    25 Sept 2026
SHARE THE FINDINGS

Share this page

DoomBench social sharing card for SpaceXAI releases Grok 4.7 for longer agentic coding tasks.
  1. DoomBench assesses “SpaceXAI releases Grok 4.7 for longer agentic coding tasks” as evidence moving toward doom, with magnitude 43 and confidence 81 out of 100 in the autonomy and agency category.

  2. The DoomBench assessment of “SpaceXAI releases Grok 4.7 for longer agentic coding tasks” is based on reporting from SpaceXAI and records the editorial rationale, source quality, attribution, and revision history.

  3. DoomBench summarizes “SpaceXAI releases Grok 4.7 for longer agentic coding tasks” as follows: SpaceXAI released Grok 4.7 through its API, Grok Build, Cursor and model routers, reporting higher coding-agent benchmark results than Grok...

    https://www.doombench.com/news/spacexai-releases-grok-4-7-for-longer-agentic-coding-tasks-2026-09-21