xAI / Grok

Grok 4.7 long-horizon agent runs impress in short bursts but fail most multi-hour tasks. Here’s exactly where it breaks, and how to scope agent work safely.

Grok Voice Transcribe 2.0 specs, pricing and API setup, plus real transcription tests against Whisper, Deepgram and AssemblyAI — and where xAI still loses

Grok for Government is live on GenAI.mil alongside ChatGPT Mil and Gemini. Here’s what DoD users get, and what it means for contractor data and compliance

Tesla is rolling out Grok in Tesla cars via OTA update. Here’s what the in-vehicle bot actually does, which models get it, and what it still can’t control

More plaintiffs joined the Grok CSAM lawsuit against xAI this week. Here’s what the multi-plaintiff case means for AI image tools and your compliance risk

Grok 4.6 is live in GitHub Copilot across VS Code, JetBrains, and Xcode. Setup steps, premium request costs, and a real repo test of where it wins

xAI’s Grok 4.6 hits 1753 LMArena Elo at roughly half the per-token cost of OpenAI and Anthropic frontier models. Here’s what that means for your routing

Grok 4.5 hits 64.7% on SWE-bench Pro while burning 4.2x fewer tokens than Opus 4.8, reframing the AI frontier race around cost-per-solved-task in 2026

Musk dissolved xAI on May 6, folding Grok into a new SpaceXAI division of SpaceX. What changed, why it matters, what it means for Grok and the IPO.

The US will test Google, Microsoft, and xAI AI models before public release, joining OpenAI and Anthropic in NIST pre-deployment review.

Google DeepMind, Microsoft, and xAI signed CAISI agreements May 5 — every major US frontier lab now does AI pre-deployment testing. What changed.