X removes 4,500 Grok-generated posts and permanently suspends 674 accounts amid Australia deepfake probe
Summary
X removes 4,500 Grok-generated posts and permanently suspends 674 accounts in the first 15 days of 2026 as Australia’s eSafety regulator investigates sexualised deepfakes.
Key Points
- Z.ai releases the 753-billion-parameter GLM 5.3 on August 14, 2026, with 40 billion active parameters per token and a 1,048,576-token context window.
- Venice Uncensored 1.2 reports a 0% refusal rate across 4,092 Dolphin-refusals prompts and 166 Nous RefusalBench prompts; its API costs $0.20 per million input tokens and $0.90 per million output tokens.
- X removes 4,500 Grok-generated items and permanently suspends more than 674 accounts between January 1 and 15, 2026, amid Australia's eSafety investigation into sexualised deepfakes.