Claude Opus 5 tops an internal writing benchmark at 2,817 Elo
RLanceMartin · x · 2026-07-26
An internal writing benchmark shared by WhatsAI puts Claude Opus 5 at #1 for editorial writing, with a score of 2817 Elo. The post says it jumped from #15 to #1 versus its predecessor when all thinking variants are considered, while keeping the same API price as Opus 4.8.
The interesting wrinkle is that effort level changes the result: at default effort, Opus 5 lands around #6; at max effort, it reaches the top spot after thinking for more than three minutes per script. The chart also shows it overtaking Claude Fable 5 and Kimi in this benchmark.
More from Models
- 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
- User Hails Unconfirmed 'DeepSeek 4.1 Flash' as an Inflection Point in LLMs — himanshustwts · 2026-09-11
- Terminal Bench v4: GLM-5.3 Leads at 41.9%, Kimi-K3 Underwhelms at 12.6% — Ok_Warning2146 · 2026-09-11
- GPT-6 Astra beats Factorio with enemies in 44 in-game hours at ~$4,500 API cost — liminal_bardo · 2026-09-11