Exact model · claude-sonnet-5

Claude Sonnet 5 API pricing and monthly cost by provider

Two exact Claude Sonnet 5 channels are visible, but APIDir does not announce a cheapest provider until a third fully comparable route passes review.

Exact model · claude-sonnet-5

Claude Sonnet 5 monthly cost comparison

Two exact Claude Sonnet 5 channels are visible, but APIDir does not announce a cheapest provider until a third fully comparable route passes review.

No complete comparison
Versioned service or route ID; family-only names are excluded.
A route without a reviewed cache meter is excluded only when this value is above zero.
No complete comparison. No provider has complete, fresh evidence for this exact workload. APIDir does not estimate missing meters.

USD-normalized monthly cost for the entered workload; every meter is per 1 million tokens.
Rank statusProviderInput priceCache-read priceOutput priceMonthly costExact configurationSources and freshness
—Evidence is aging or staleAnthropic APIofficial$2 / 1M input tokens$0.2 / 1M cached input tokens$10 / 1M output tokensUnavailableEvidence is aging or staleclaude-sonnet-5 · Anthropic Messages API · standard
  • Input · checked · fresh through staleFirst-party source
  • Cached input · checked · fresh through staleFirst-party source
  • Output · checked · fresh through staleFirst-party source
Row eligibility uses only meters with non-zero workload.
—Evidence is aging or staleOpenRouterrouter$2 / 1M input tokens$0.2 / 1M cached input tokens$10 / 1M output tokensUnavailableEvidence is aging or staleanthropic/claude-sonnet-5 · router standard
  • Input · checked · fresh through staleFirst-party source
  • Cached input · checked · fresh through staleFirst-party source
  • Output · checked · fresh through staleFirst-party source
Row eligibility uses only meters with non-zero workload.
Missing and excluded providers (6)
  • Anthropic API: Evidence is aging or stale.
  • OpenRouter: Evidence is aging or stale.
  • Together AI: No accepted exact Claude Sonnet 5 route and price observation was found in this review.
  • Fireworks AI: No accepted exact Claude Sonnet 5 route and price observation was found in this review.
  • DeepInfra: No accepted exact Claude Sonnet 5 route and price observation was found in this review.
  • Novita AI: No accepted exact Claude Sonnet 5 route and price observation was found in this review.

Strict boundary: Standard rates and cache reads are included. Cache writes, batch discounts, long-context premiums, tools, latency, rate limits, support, taxes, and service quality are outside this comparison.

Embeddable result

Put this exact-model comparison on your site

The SVG card is cookie-free and links back to the same source-auditable workload. It shows a rank only while at least three routes qualify.

<script async referrerpolicy="no-referrer" src="https://apidir.dev/embed/model-cost.js?model=claude-api&input=100000000&cached=0&output=20000000"></script>

One exact AI model compared across three provider routes with separate token meters and an incomplete evidence route
AI concept diagramConcept diagram: exact identity and complete meters decide eligibility. An incomplete route stays visible as a gap and cannot become a winner.

Exact-model alert · controlled rollout

Join the Claude Sonnet 5 price alert waitlist.

Confirm your address to join the waitlist. Delivery remains paused until sender, suppression, and evidence checks are ready; ambiguous or stale changes stay quarantined.

Evidence boundary

What the monthly comparison proves

Standard rates and cache reads are included. Cache writes, batch discounts, long-context premiums, tools, latency, rate limits, support, taxes, and service quality are outside this comparison.

A row qualifies only when the exact model route, channel, region, mode, tier, and input/output meters match the reviewed definition. Cached input becomes mandatory when your cached-input volume is above zero.

Questions developers ask

Which exact Claude Sonnet 5 route is compared?

This page is locked to claude-sonnet-5. Family-only names, near matches, different model versions, and unreviewed aliases are excluded.

When is a lowest-cost label shown?

Only when at least three unique providers have complete fresh evidence for every token meter used by the submitted workload. Equal monthly costs share a rank.

What is excluded from monthly cost?

Retries, failed requests, networking, storage, taxes, support, dedicated capacity, latency, rate limits, and provider-specific service quality are outside this token-meter estimate.