Getting started
Pricing
EzRouter charges separately for input, output, cache-read, and cache-write tokens. There is no subscription or minimum spend. All rates below are in US dollars per 1 million tokens and reflect the enabled EzRouter production catalog on July 11, 2026.
The dashboard model catalog and usage logs are the final source of truth for availability and billing. Query GET /v1/models before hard-coding a model ID because routes can be added, disabled, or temporarily unavailable.
Current model rates
| Model ID | Input | Output | Cache read | Cache write |
|---|---|---|---|---|
claude-haiku-4-5 | $0.60 | $3.00 | $0.06 | $0.75 |
claude-opus-4-6 | $3.00 | $15.00 | $0.30 | $3.75 |
claude-opus-4-7 | $3.00 | $15.00 | $0.30 | $3.75 |
claude-opus-4-8 | $3.00 | $15.00 | $0.30 | $3.75 |
claude-sonnet-4-6 | $1.80 | $9.00 | $0.18 | $2.25 |
claude-sonnet-5 | $1.80 | $9.00 | $0.18 | $2.25 |
claude-fable-5 | $6.00 | $30.00 | $0.60 | $7.50 |
gpt-5.4 | $0.50 | $3.00 | $0.05 | $0.50 |
gpt-5.4-mini | $0.15 | $0.90 | $0.015 | $0.15 |
gpt-5.5 | $1.00 | $6.00 | $0.10 | $1.00 |
gpt-5.6-luna | $0.20 | $1.20 | $0.02 | $0.25 |
gpt-5.6-sol | $1.00 | $6.00 | $0.10 | $1.25 |
gpt-5.6-terra | $0.50 | $3.00 | $0.05 | $0.625 |
kimi-k2.7-code | $0.95 | $4.00 | $0.19 | $0.95 |
deepseek-v4-flash | $0.14 | $0.28 | $0.0028 | $0.14 |
deepseek-v4-pro | $0.45 | $0.90 | $0.00375 | $0.45 |
glm-5.2 | $1.40 | $4.40 | $0.26 | $0.00 |
MiniMax-M3 | $1.20 | $4.80 | $0.24 | $0.24 |
gpt-image-2 | $2.50 | $15.00 | $1.00 | $3.125 |
gpt-image-2 exposes token-rate fields in the model catalog, but image generation can include image-specific usage accounting. Use the charge shown in the dashboard usage log for final reconciliation.
How billing is calculated
For a request without cache usage:
charge = (input_tokens / 1,000,000) * input_rate
+ (output_tokens / 1,000,000) * output_rateWhen the response reports cache usage, cached input and newly written cache tokens use the corresponding cache rates instead of the regular input rate. Cache behavior depends on the model and upstream route; a listed cache rate does not guarantee that every request will create or hit a cache entry.
Inspect the response usage object and the dashboard usage log when reconciling a request. Relevant fields can include prompt_tokens_details.cached_tokens, claude_cache_creation_5_m_tokens, and claude_cache_creation_1_h_tokens, depending on the API surface and model.
Price and availability changes
EzRouter may update a model rate or disable a route as supplier cost and capacity change. The homepage pricing calculator loads the same dated site snapshot used for this page, while the authenticated dashboard reflects the current catalog and final charged amount.
Top up based on recent usage rather than assuming a model or rate will remain unchanged indefinitely. Contact [email protected] for billing questions.