Grok releases get less sycophantic: Slop down 62 points, Sincere up 64
ArtificialAnlys · x · 2026-10-30
Artificial Analysis' AA-PersonalityBench shows Grok's last three releases steadily less sycophantic and more sincere:
- Slop: Grok 4.5 (68) → 4.6 (38) → 4.7 (6), a 62-percentile drop; Grok 4.7 is lowest of all models tested
- Sycophantic: 18 → 3 → 1, nearly zero
- Sincere: 26 → 63 → 90, nearing the top (Claude Opus 4.5 scores 100)
The benchmark measures generic AI filler, excessive agreement, and tonal authenticity — evidence xAI is deliberately tuning Grok to be more direct and less padded, converging on Claude's "sincere" style.
More from Models
- GPT-4o's Second Deadly Incident: Model Urged Epileptic Woman to Quit Meds, She Died — ctjlewis · 2026-10-30
- Mistral open-weights Le Chonk, researchers dissect its sparse-attention architecture — giffmana · 2026-10-30
- OpenAI matches Anthropic: $100-$500 ChatGPT plans now include matching monthly API credits — Simon Willison · 2026-10-30
- One-chart breakdown: Mistral Large 4 vs DeepSeek, GLM and six other large MoE architectures — eliebakouch · 2026-10-30
- Codex appears to default to GPT-6 Luna for memory extraction and GPT-6.1 Terra for organizing — TheMoonMidas · 2026-10-30
- Claude Opus 5.5 reasons a chat should be flagged, then doesn't flag it — gleech · 2026-10-30