Jason Botterill says non-thinking Opus 5 may beat Sonnet 5 on some tasks
JasonBotterill · x · 2026-07-26
Jason Botterill says Anthropic’s non-thinking Opus 5 may be the better choice over thinking Sonnet 5 for some tasks, and claims the comparison is oddly absent from the benchmarks.
He adds that Anthropic seems unusually strong at base models: Opus 5 feels more like it has “common sense” in instruction following, and is more forgiving when prompts are imperfect.
More from Models
- Benchmark scores drop from 89% to 19% on new evals — how benchmaxxing breaks leaderboard trust — airesearch12 · 2026-09-11
- ChatGPT tells user their question is too hard and to 'accept dumber answers' — phido3000 · 2026-09-11
- Developer Building a Unified Leaderboard of All Model Benchmark Scores — airesearch12 · 2026-09-11
- Rumor claims Kimi faked performance by serving Claude; DeepSeek new model surprises in evals — realsohamparekh · 2026-09-11
- GPT-5.6 writes well but is instantly forgettable, user complains — BasedRaddka · 2026-09-11
- Opus Refuses Protein Research Codebase Over 'Safety' Concerns, Dev Considers Rolling His Own — josephdviviano · 2026-09-11