GPT-6-Astra Tops HalluHard Hallucination Benchmark

GPT-6-Astra takes the top spot on the HalluHard hallucination benchmark, significantly beating all rivals including Fable 5 with and without web search. The multi-turn benchmark also shows frontier models still hallucinate heavily in long conversations.

2026-09-18 ~ 2026-09-18 · 4 related posts

• HalluHard Benchmark: GPT-6-Astra Clearly Ahead of Fable 5 on Hallucinations — maksym_andr · 2026-09-18
• HalluHard Benchmark Shows Frontier Models Still Hallucinate Heavily in Multi-Turn Tasks — maksym_andr · 2026-09-18

2 near-duplicate retellings: maksym_andr · maksym_andr