Start with Opus 5.5
A strong default for sophisticated agentic coding and knowledge work when quality matters more than minimum cost.
Model guide
Pick the right Claude model for the job.
All current models support text and image input, text output, multilingual work, vision, and tool use. The practical differences are capability, speed, and token pricing.
| Model | Best for | Context | Max output | Input / MTok | Output / MTok |
|---|---|---|---|---|---|
| Claude Fable 5.1 | Demanding reasoning, long-horizon agents | 1M | 128K | $10 | $50 |
| Claude Opus 5.5 | Long-running agentic coding, knowledge work | 1M | 128K | $4 | $20 |
| Claude Sonnet 5.5 | General coding, automation, production apps | 1M | 128K | $2 | $10 |
| Claude Haiku 5.5 | High-volume, low-latency tasks | 1M | 128K | From $0.10 | From $0.50 |
Pricing shown is the public Anthropic token pricing currently used by TokenRealm's metering logic. Haiku pricing can vary for prompts above 100K tokens.
Selection guide
Which one should you start with?
Choose Sonnet 5.5 for balance
Well suited to everyday coding, developer tools, automation, chat products, and high-quality general workloads.
Use Haiku 5.5 for scale
Best for fast classification, extraction, routing, structured data work, and other latency-sensitive workloads.
Move to Fable 5.1 when needed
Designed for demanding reasoning and long-horizon agentic work where other models do not meet your evaluation target.
Model aliases
Simple names for stable integrations.
TokenRealm supports short aliases so clients can request the current model in each family without embedding long IDs everywhere.
fable→ claude-fable-5-1opus→ claude-opus-5-5sonnet→ claude-sonnet-5-5haiku→ claude-haiku-5-5