The switch itself is one field: set model to claude-fable-5-2.
Run your own test prompts first, since Anthropic changed how often 5.2 refuses.
Fast mode doubles the price; turn it on only where latency matters.
Claude Fable 5.2 costs 20% less per token than 5.1 on WayJet ($8 / $40 per million tokens, against $10 / $50), and Anthropic says it is up to 30% faster. Switching is easy. Switching without surprises takes five minutes more.
1. Change one field
WayJet's API is OpenAI-compatible, so the only change is the model id:
{
"model": "claude-fable-5-2",
"messages": [{ "role": "user", "content": "Summarise this ticket in two lines." }]
}
Your key, endpoint and code stay the same.
2. Run your test set on both
Before moving production traffic, send the same 20–50 real prompts to claude-fable-5-1 and claude-fable-5-2 and compare the answers. Anthropic says 5.2 blocks about 55% fewer requests unnecessarily. That is usually good news, but if your product relies on the model declining certain requests, check those cases.
3. Decide on fast mode
Fable 5.2 has an optional fast mode at double the price. Use it for interactive features where a person is waiting, and leave background jobs on the standard rate.
4. Compare the spend
After a day on the new model, open Usage in the WayJet dashboard and filter by model. You should see the cost per request fall by about a fifth for the same prompts; if output got longer, the saving will be smaller.
5. Keep 5.1 as a fallback for a week
Leave the old id in your config so you can switch back in seconds if something in your test set missed a case.
Источники и контекст
- SD Times: Anthropic introduces Claude Fable 5.2, Mythos 5.2sdtimes.com · Проверено 12 нояб. 2026 г.
- WayJet model catalogwayjet.ai · Проверено 12 нояб. 2026 г.

