Published 10 hours ago • loading... • Updated 20 minutes ago
Anthropic Unveils Claude Opus 5.5
Anthropic says the model matches top-tier performance on most tasks while cutting run costs 40% and improving safety against prompt injection.
On Tuesday, Anthropic released Claude Opus 5.5 while OpenAI launched GPT-6 Sol and GPT-6 Luna for ChatGPT Work and Codex users, intensifying competition for developer market share through simultaneous product rollouts.
Cheaper open-weight models from Alibaba, Moonshot AI, and DeepSeek pressure both labs to commoditize flagship capabilities. This market shift forces leaders to cut costs while maintaining performance for enterprise customers.
Anthropic's Opus 5.5 costs 40% less than Opus 5 at $4 input and $20 output tokens per million. OpenAI reduced Sol and Luna pricing by 50% through infrastructure optimizations and prompt caching improvements.
Following Anthropic CEO Dario Amodei's call to 'pace the frontier,' safety remains central to the release. Opus 5.5 underwent external testing by METR and Frontier Design before launch, with safeguards for cybersecurity and biology.
Anthropic plans to release Claude Sonnet 5.5 and Claude Haiku 5.5 in coming weeks, expanding its Claude 5.5 family. Rapid-fire releases occur as both labs prepare for potential initial public offerings, aiming to solidify valuations.
Daily Technology, Business and Entrepreneur podcast
Daily Technology, Business and Entrepreneur podcast
Harvey’s Margins, Are Insects Worth More Than Humans?, Meta Tests Human Help for Muse | Diet TBPN
TBPN Live discuss Anthropic’s Claude Opus 5.5 launch, highlighting “pacing the frontier,” cheaper pricing, faster outputs, and Fable 5.1–class capabilities