2h ago

Claude Opus 4.8 Reduces Hallucinations On Missing-Context Benchmarks Versus 4.7

——0——
Original post
OPwh#1430wh@NREHIEW_

A big problem with Opus 4.7 was its hallucinations. Good to see slight improvements with Opus 4.8 at least on the benchmarks they report on

10:18 AM · May 28, 2026 View on X
REPLYwh#1430wh@NREHIEW_

Interestingly, a significant amount of hallucination evals used tests the models ability to resist pressure from incorrect info supplied by the user.

I'm not entirely sure if thats reflective of the hallucinations most people encounter

whwh@nrehiew_

A big problem with Opus 4.7 was its hallucinations. Good to see slight improvements with Opus 4.8 at least on the benchmarks they report on

5:18 PM · May 28, 2026 · 1.3K Views
5:18 PM · May 28, 2026 · 344 Views