Anthropic: Claude Fable 5.2

• SandbaseServed when a request names it.Cache $0.20Input $8.00Cache read $0.20Output $40.00Latency —Speed —Uptime —$8.00 in$40.00 out

  Price history
• Amazon BedrockServed when a request names it.Cache $0.21 · 5.37 s · 84 tok/s5.37 s · 84 tok/sInput $8.40Cache read $0.21Output $42.00Latency 5.37 sSpeed 84 tok/sUptime —$8.40 in$42.00 out

  1M context · 128K max output

• AnthropicServed when a request names it.Cache $0.21 · 4.46 s · 78 tok/s · 100 %4.46 s · 78 tok/sInput $8.40Cache read $0.21Output $42.00Latency 4.46 sSpeed 78 tok/sUptime 100 %$8.40 in$42.00 out
  • Route 1Served when a request names it.1M context · 128K max outputCache $0.21 · 4.46 s · 78 tok/s · 100 %4.46 s · 78 tok/sInput $8.40Cache read $0.21Output $42.00Latency 4.46 sSpeed 78 tok/sUptime 100 %$8.40 in$42.00 out
  • Route 21M context · 128K max outputCache $0.42 · 2.69 s · 166 tok/s · 100 %2.69 s · 166 tok/sInput $16.80Cache read $0.42Output $84.00Latency 2.69 sSpeed 166 tok/sUptime 100 %$16.80 in$84.00 out
• AzureServed when a request names it.Cache $0.21 · 5.03 s · 71 tok/s · 100 %5.03 s · 71 tok/sInput $8.40Cache read $0.21Output $42.00Latency 5.03 sSpeed 71 tok/sUptime 100 %$8.40 in$42.00 out

  1M context · 128K max output

• GoogleServed when a request names it.Cache $0.21 · 3.96 s · 79 tok/s · 100 %3.96 s · 79 tok/sInput $8.40Cache read $0.21Output $42.00Latency 3.96 sSpeed 79 tok/sUptime 100 %$8.40 in$42.00 out

  1M context · 128K max output

• QubaxServed when a request names it.Cache $0.21Input $8.40Cache read $0.21Output $42.00Latency —Speed —Uptime —$8.40 in$42.00 out

  1M context · 128K max output

  Price history
• SurplusRequests go hereCache $0.24Input $9.60Cache read $0.24Output $48.00Latency —Speed —Uptime —$9.60 in$48.00 out

  1M context · 128K max output

  Price history
• NanoGPTCache $0.25Input $10.00Cache read $0.25Output $50.00Latency —Speed —Uptime —$10.00 in$50.00 out

  1M context · 128K max output

  Price history

Per 1M tokens: the provider's own price. Latency is the median time to the first token; speed is the median output rate; uptime covers the last day (on a wide table, three bars: the last day, 30 minutes and 5 minutes). “Requests go here” marks where a request that names no provider goes now. Logos are trademarks of their owners.