1. Smartest non-reasoning model

• Highest intelligence among models without a reasoning step.
• Great for tasks where speed + accuracy matter without deep chain-of-thought.

2. Excellent instruction following

• Very strong at structured tasks, formatting, and precise execution.
• Ideal for productized workflows and deterministic outputs.

3. Reliable tool calling

• Works smoothly with Web Search, File Search, Image Generation, and Code Interpreter.
• Supports MCP and advanced tool-enabled API flows.

4. Large 1M-token context window

• Allows extremely long conversations, large documents, and multi-file use cases.
• Handles context-heavy tasks without requiring chunking.

5. Low latency (no reasoning step)

• Faster responses than GPT-5 family when reasoning mode isn't required.
• More predictable timing for production use.

6. Multimodal input

• Accepts text + image.
• Output is text only.

7. Supports fine-tuning

• Can be fine-tuned for specialized tasks.
• Also supports distillation for smaller custom models.

• Highest capability in the family: described as “our most powerful model yet” by Anthropic.
• Exceptional at long-running tasks requiring thousands of steps and sustained focus (e.g., continuous codebase work for hours).
• Excellent performance on benchmarks: e.g., SWE-bench 72.5 % and Terminal-bench 43.2 %.
• Designed for complex agentic workflows, deep reasoning, tool use, and large context windows.
• Placed under a higher safety classification (ASL-3) due to its frontier capability and risk profile.