Anthropic Launches Claude Opus 5: Near-Frontier Performance at Half the Price
Summary: Anthropic's Claude Opus 5 delivers close-to-frontier reasoning at $5/$25 per million input/output tokens — identical to its predecessor's price — while setting new state-of-the-art marks on coding and knowledge-work benchmarks.
Key Facts
- Pricing held flat at $5/M input, $25/M output tokens (same as Opus 4.8), yet outscores prior models on Frontier-Bench and GDPval-AA evaluations
- Effort setting: users choose low, medium, or high reasoning depth, letting applications tune cost vs. capability at the API level
- Markedly improved self-verification — the model retries and corrects mistakes without prompting, reducing iteration rounds
- Available immediately on Claude.ai, the Claude API, Claude Code, and Claude Max subscriptions
Why It Matters
Opus 5 cements a capable "sub-flagship" tier just below Fable 5's price point, giving enterprise developers a credible default for demanding workloads without committing to Fable 5 costs. The effort setting is notable as the first native mechanism for developers to programmatically trade reasoning depth for latency and cost within a single model call.
Read More
- Anthropic launches Claude Opus 5 with efficiency, safety improvements — SiliconANGLE
- Anthropic launches Opus 5 — TechCrunch
How this story unfolded
- Google Launches Three Gemini Models at Once: 3.6 Flash, 3.5 Flash-Lite, 3.5 Flash Cyber
- GPT-5.6's Prediction Window Opens: 1.5M Token Context and Alignment Fix Expected
- GPT-5.6 Launch This Week Has 83% Odds, OpenAI Scientist Confirms 'Meaningful Leap'
- Gemini 3.5 Pro Still Unreleased With 11 Days Left in Google's June Commitment Window
Tools in this story
- LLM
- — Large Language Model의 약자로, '거대 언어 모델'이라고 해요. ChatGPT, Claude 같은 AI가 바로 LLM이에요. 엄청나게 많은 텍스트를 학습해서 사람처럼 글을 쓰고 대화할 수 있어요.