Claude Opus 5 Details: Near-Frontier at Half Price with Proactive Vision Skills
Simon Willison · rss · 2026-07-25
Anthropic has released the new model Claude Opus 5, claiming to approach the frontier intelligence of Claude Fable 5 at half the price while being thoughtful and proactive. It currently leads the Artificial Analysis leaderboard.
Priced identically to its predecessor Opus 4.8, it retains a twice-the-cost "fast mode." Opus 5 demonstrates extreme proactiveness: in a test where it was intentionally given no way to view an image, it wrote its own computer vision pipeline to extract geometry from raw pixels and reconstructed a 3D machine part.
Regarding safety, Opus 5 is better at finding vulnerabilities but has deliberately not been trained on how to exploit them, reducing cyber threats. Anthropic also published an accompanying prompting guide.
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