Release and model choices
Fable is the broadly available variant. Organizations seeking Mythos access apply through Anthropic's verification programs; selecting Mythos in this local counter does not authorize or enable model access.
The announcement highlights improvements in coding, reasoning, vision and scientific and engineering tasks. Anthropic still recommends Opus 5.5 as a starting point for most everyday coding and knowledge work, with Fable or Mythos for harder problems.
| Model | Local counting | Open tool |
|---|---|---|
| Claude Fable 5.2 | Local estimate | Select Claude Fable 5.2 |
| Claude Mythos 5.2 | Local estimate | Select Claude Mythos 5.2 |
How to count Claude Fable 5.2 & Mythos 5.2 tokens
English planning rule: characters ÷ 3.5 × 1.3, rounded up. This is the provider's general English-oriented planning guidance, not a published Claude Fable 5.2 & Mythos 5.2 tokenizer. Other languages, code, emoji and punctuation can differ substantially. The estimate does not produce model token IDs or guarantee a context or billing limit.
Paste text into Plain text mode, or use Chat messages to count each message body independently. Switching models preserves your input. Chat mode excludes role names, separators and provider wrappers; adding bodies does not reproduce a serialized API conversation.
Local processing does not send your prompt to a model API or generate a response. Published tokenizer assets may be loaded by the browser, while the text analysis itself runs on your device.
Request limits and usage
Anthropic's token-counting endpoint provides a preflight estimate that can differ from actual Messages usage. This site's English character heuristic is less precise still and cannot measure safeguard-related processing or complete request formatting.
For a production request, include the actual conversation, system instructions, tool schemas and supported media in the provider's counting workflow when available. Check response usage after generation and account for subsequent calls, cache behavior, reasoning and output separately.
- A model's advertised context window is not a guaranteed amount of pasted text.
- A faster model or a cheaper token rate does not establish a lower total bill for every task.
- Recheck the exact model version and counting method when migrating an existing prompt.
Tiktokenizer Reference Sources
Tokenizer behavior and model limits change. Verify production decisions with current provider documentation: