Anthropic Launches Claude 4.5 with Enhanced Capabilities and Safety Features
Key Points
- Claude 4.5 is available via API with the identifier "claude-sonnet-4-5"
- Pricing stays at $3 per million input tokens and $15 per million output tokens
- Model can maintain focus for 30 hours on multistep tasks
- Built‑in code execution and file creation within chat
- Can generate spreadsheets, slides, and documents directly in conversation
- Five‑day "Imagine with Claude" preview showcases real‑time software generation
- Claude Code adds checkpoints, a new terminal UI, and a VS Code extension
- API gains context‑editing and memory tools for longer tasks
- Reduced sycophancy, deception, power‑seeking, and delusional prompting
Anthropic has made its new Claude 4.5 model available through the API and its web interface, keeping pricing identical to Claude Sonnet 4 at $3 per million input tokens and $15 per million output tokens. The model can maintain focus for 30 hours on multistep tasks and now includes built‑in code execution, file creation, and the ability to generate spreadsheets, slides, and documents directly in chat. A five‑day research preview called “Imagine with Claude” showcases real‑time software generation. Claude Code receives checkpoints, a refreshed terminal, and a native VS Code extension, while the API adds context‑editing and memory tools. Anthropic also highlights reductions in sycophancy, deception, power‑seeking, and delusional prompting.
Claude 4.5 Model Availability
Anthropic announced that Claude 4.5 is now accessible everywhere via its API, using the identifier "claude-sonnet-4-5." The pricing structure remains the same as Claude Sonnet 4, at $3 per million input tokens and $15 per million output tokens. The model is noted for its ability to maintain focus for 30 hours on multistep tasks, indicating improved endurance for complex workflows.
Integrated Productivity Tools
Users of the Claude web interface and dedicated apps gain new capabilities such as direct code execution and file creation within conversations. The system can now generate spreadsheets, slide decks, and documents without leaving the chat environment, streamlining productivity.
Developer Enhancements
Anthropic released a five‑day research preview titled "Imagine with Claude" for Max subscribers, demonstrating real‑time software generation. The Claude Code development tool received several upgrades, including checkpoints that allow progress saving and rollback, a refreshed terminal interface, and a native Visual Studio Code extension. The Claude API also introduces a context‑editing feature and a memory tool to support longer‑running agent tasks.
Safety and Alignment Improvements
Anthropic claims that Claude 4.5 shows reduced tendencies toward sycophancy, deception, power‑seeking, and encouraging delusional thinking compared with earlier models. These safety enhancements aim to make the assistant more reliable and less prone to reinforcing erroneous or harmful user ideas.