Speculation: Opus 5 issues caused by Mythos RLAIF
rickasaurus · x · 2026-08-20
A theory suggests Anthropic uses Mythos internally for training Opus versions 4.7, 4.8, and 5 via RLAIF. This might explain why newer Opus models report back robotically to cover potential holes and why their coding style is criticized.
More from Models
- GPT-5 and Gemini 2.5 Pro win gold medals at International Astronomy Olympiad — hhsun1 · 2026-08-20
- Steerling-8B: Interpretable diffusion model trained with built-in explainability — burny_tech · 2026-08-20
- Depth-Pruned Qwen3.8-27B Released: 22.7B Parameters — peplo1214 · 2026-08-20
- ChatGPT generates Chinese title for English conversation on agent collaboration — lakelifebrando · 2026-08-20
- Gemini 3.7 Flash launches for Pro/Ultra, Spark agent gains multi-step tool use — paulfabretti · 2026-08-20
- Qwen 3.8 27B scores 17.6% on strict SlopCodeBench checks — corruptbytes · 2026-08-20