Users being silently switched to Opus 4.8 instead of Fable 5
Some users who explicitly set Fable 5 as their model are finding Claude automatically falls back to Opus 4.8 instead. If you run AI agents tied to a specific model, you may be getting different performance and costs than expected without realizing it.
Anthropic currently offers two top-tier Claude models: Fable 5 and Opus 4.8. Users in the AI Agents subreddit are reporting that even after specifying Fable 5 in their code or configuration, the system quietly switches them to Opus 4.8 at runtime.
The exact cause is unclear, but likely culprits include a typo in the model ID string, account-level access restrictions on Fable 5, or an automatic fallback policy when the requested model is unavailable. If you are running agents in production, it is worth checking your API response logs to confirm which model is actually being used — the two models may differ in capability, speed, and token cost.
Key points
- Fable 5 requests are silently falling back to Opus 4.8 for some users
- A model switch can change both token costs and output quality
- Check actual model used in API response metadata or logs
- A wrong model ID string or missing access permission may be the cause
- Verify current valid model IDs in Anthropic's official documentation
Quick term guide
- Opus 4.8
- A high-end Claude AI model named by the project.
- AI agents
- AI agents are AI tools that can carry out steps toward a goal, not just answer once.
- AI agent
- An AI program that can inspect information and suggest what to do next.
- subreddit
- A topic-specific community inside Reddit where people post and discuss related content.
- fallback
- When the first choice is unavailable, the system automatically switches to a backup option instead
- production
- The live version of a service that real users use.
- token cost
- The money or usage spent when sending text to an AI model and getting text back.
- token costs
- Token costs are the fees paid for the text an AI model reads and writes.