Anthropic launches Claude Opus 5 at half Fable 5's price

What happened: Anthropic launched Claude Opus 5 on 24 July 2026. It's now the default model for Claude Max subscribers and the top-tier option on Claude Pro.

The numbers: Pricing holds at $5 per million input tokens and $25 per million output tokens, matching Opus 4.8 but roughly half what Claude Fable 5 costs. It carries a 1-million-token context window and up to 128K output tokens. Opus 5 beats Fable 5 on 8 of 13 benchmarks, including a 9.7-point jump on Frontier-Bench (43.3%) and an 8.5-point gain on AutomationBench.

Where it's available: Claude, Claude Code, the Claude API (model ID claude-opus-5), Amazon Bedrock, Google Cloud Vertex AI and Microsoft Foundry.

Why it matters: Anthropic is pitching Opus 5 as near-Fable 5 intelligence at a steep discount, while also calling it its safest model yet. It scored zero on Anthropic's internal benchmark for exploiting vulnerabilities and triggers safety guardrails 85% less often than Fable 5.

The details: A new low/medium/high effort setting lets developers trade cost for capability per request. The API now supports interrupting a model mid-workflow to change its tools, and prompts blocked by safety filters can be rerouted to a different model. The model can also check its own work — in one test it wrote a "computer vision pipeline to pull the geometry from the raw pixels."

The context: Opus 5 is Anthropic's fourth model release in under two months, following Mythos 5, Fable 5 and Sonnet 5 in June. A checkpoint codenamed "Claude Honeycomb EAP" briefly surfaced in Cursor's model picker on 9 July, ahead of the official launch.

What's next: With pricing held steady and safety metrics improved, Anthropic is betting Opus 5 becomes the default choice for developers who found Fable 5 too costly for everyday use. Details via SiliconANGLE.