API credits + priority limits
Anthropic Startup Program
Compare 8 Anthropic models in the TokenCOGS registry, then run the same workload through the cost calculators.
Start with the headline price range, then use the full table for the complete lineup.
USD per 1M text tokens. Live pricing is checked against the central TokenCOGS registry when this page loads.
| Model | Input | Output | Cached input | Context | |
|---|---|---|---|---|---|
| Claude Fable 5.1 | $10 | $50 | $0.25 | Not listed in registry | Details → |
| Claude Fable 5Legacy / still live | $10 | $50 | $1 | Not listed in registry | Calculate → |
| Claude Opus 5 | $5 | $25 | $0.5 | 1M tokens | Details → |
| Claude Opus 4.8 | $5 | $25 | $0.5 | Not listed in registry | Calculate → |
| Claude Sonnet 5 | $2 | $10 | $0.2 | 1M tokens | Details → |
| Claude Sonnet 4.6 | $3 | $15 | $0.3 | Not listed in registry | Calculate → |
| Claude Sonnet 4.5 | $3 | $15 | $0.3 | Not listed in registry | Calculate → |
| Claude Haiku 4.5 | $1 | $5 | $0.1 | 200K tokens | Details → |
Anthropic Startup Program
Use this table to eliminate models that do not fit your unit economics, then benchmark the remaining candidates on your own task for output quality, latency, rate limits and operational reliability.
Claude Haiku 4.5 currently has the lowest tracked input-token price in this Anthropic set. That does not automatically make it the lowest-cost model for every workload because output ratios, caching, long-context rules and tool usage can change the result.
Use this provider’s price table as a starting point, then estimate the cost of your actual workload, users or multi-step tasks.