Gemini 4 Argon Posts 15% Hallucination Rate, Far Below GPT-6 Astra's 45%
i_dg23 · x · 2026-10-01
- Artificial Analysis numbers show Gemini 4 Argon with a 15% hallucination rate vs Grok 4.7 at 29%, GPT-6 Astra at 45%, Opus 5.5 at 59%, and Fable 5.1 at 69%.
- On accuracy Argon (50%) still trails Opus 5.5 max (66%), but it says "I don't know" instead of making things up.
- Third-party relay of benchmark data; hands-on validation pending.
More from Models
- One-line take: GPT-6.1 Sol is underrated, says AI commentator — mallow610 · 2026-10-01
- "Opus 5.5 is the cheapest model" — human hours saved beat token pricing — CamBrazy3 · 2026-10-01
- CUHK study maps when recurrence helps in looped language models, proposes history-state injection — CUHK-CSE · 2026-10-01
- Local 27B Face-off: Dirk-Qwen3.8 Beats Swift-1.5 on a 200-Question Personal Eval — norenEnmotalen · 2026-10-01
- Influencers hype Gemini 4 Argon: GOATED or hopelessly benchmaxed? — thatroblennon · 2026-10-01
- Reddit Pushback: OpenAI Users Subsidize Failed Experiments Like Atlas and Sora Via Price Hikes — dagerika · 2026-10-01