Skip to content

Claude ​

Claude Max Pool — 0.24x of official Anthropic pricing. Full capacity. Supports claude-fable-5-1, claude-opus-5-5, claude-opus-5, claude-opus-4-8, claude-opus-4-7, claude-opus-4-6, claude-sonnet-5, claude-sonnet-4-6, and claude-haiku-4-5. Quality and behavior are identical to official Claude Max. Currently limited to Claude Code only.

Claude via Kiro — 0.08x of official Anthropic pricing. Partial capacity. Supports all models above except claude-fable-5-1. Kiro routes may have some system prompt contamination.

Most Kiro providers on the market use free-tier accounts with significantly lower cache and intelligence. CE uses enterprise accounts to ensure higher quality.


Official Anthropic Pricing ​

Official Anthropic pricing for reference. Per 1M tokens, USD.

ModelContextCache Read5m Cache Write1h Cache WriteInputOutput
claude-fable-5-11M$0.25$12.50$20.00$10.00$50.00
claude-opus-5-51M$0.20$5.00$8.00$4.00$20.00
claude-opus-51M$0.50$6.25$10.00$5.00$25.00
claude-opus-4-81M$0.50$6.25$10.00$5.00$25.00
claude-opus-4-71M$0.50$6.25$10.00$5.00$25.00
claude-opus-4-61M$0.50$6.25$10.00$5.00$25.00
claude-sonnet-51M$0.20$2.50$4.00$2.00$10.00
claude-sonnet-4-61M$0.30$3.75$6.00$3.00$15.00
claude-haiku-4-5200K$0.10$1.25$2.00$1.00$5.00

CE Max Pool Pricing (0.24x) ​

ModelCache ReadInputOutput
claude-fable-5-1$0.06$2.40$12.00
claude-opus-5-5$0.048$0.96$4.80
claude-opus-5$0.12$1.20$6.00
claude-opus-4-8$0.12$1.20$6.00
claude-opus-4-7$0.12$1.20$6.00
claude-opus-4-6$0.12$1.20$6.00
claude-sonnet-5$0.048$0.48$2.40
claude-sonnet-4-6$0.072$0.72$3.60
claude-haiku-4-5$0.024$0.24$1.20

CE Kiro Pricing (0.08x) ​

ModelCache ReadInputOutput
claude-opus-5-5$0.016$0.32$1.60
claude-opus-5$0.04$0.40$2.00
claude-opus-4-8$0.04$0.40$2.00
claude-opus-4-7$0.04$0.40$2.00
claude-opus-4-6$0.04$0.40$2.00
claude-sonnet-5$0.016$0.16$0.80
claude-sonnet-4-6$0.024$0.24$1.20
claude-haiku-4-5$0.008$0.08$0.40

Notes ​

Cache writes are billed at 1.25x input (5-minute TTL) or 2x input (1-hour TTL). These multipliers apply to both Max Pool and Kiro at their respective rates.

Fable 5.1 cache reads are billed at 0.025x input, and Opus 5.5 cache reads at 0.05x input, following official Anthropic pricing.

Claude 4.7 and later models use a newer tokenizer that produces approximately 30% more tokens for the same text compared to earlier models. The exact increase depends on the content and workload.

Opus 4.6 and Sonnet 4.6 are better suited for writing and conversation tasks.