AI API 定價與模型 | Modelflare

比較目前可用的模型、模型群組與價格和官方 API 定價，並查看相容的 API 格式。

Prices are in USD. The listed group rate applies before any request-specific multiplier. Your account may have different access or discounts.

模型群組條件單位價格
MiniMax-H3Media輸出 · 2k每秒USD 0.1521
claude-fable-5claude-award輸入 · 標準每 1M TokenUSD 2.5
claude-fable-5claude-first-topup輸入 · 標準每 1M TokenUSD 0.3
claude-fable-5claude-premium輸入 · 標準每 1M TokenUSD 5
claude-fable-5claude-stable輸入 · 標準每 1M TokenUSD 3.15
claude-fable-5-1claude-award輸入 · 標準每 1M TokenUSD 2.5
claude-fable-5-1claude-first-topup輸入 · 標準每 1M TokenUSD 0.3
claude-fable-5-1claude-premium輸入 · 標準每 1M TokenUSD 5
claude-fable-5-1claude-stable輸入 · 標準每 1M TokenUSD 3.15
claude-haiku-4-5claude-award輸入 · 標準每 1M TokenUSD 0.25
claude-haiku-4-5claude-first-topup輸入 · 標準每 1M TokenUSD 0.03
claude-haiku-4-5claude-premium輸入 · 標準每 1M TokenUSD 0.5
claude-haiku-4-5claude-stable輸入 · 標準每 1M TokenUSD 0.315
claude-haiku-5-5claude-award輸入 · Context length <= 100000每 1M TokenUSD 0.025
claude-haiku-5-5claude-premium輸入 · Context length <= 100000每 1M TokenUSD 0.05
claude-haiku-5-5claude-stable輸入 · Context length <= 100000每 1M TokenUSD 0.0315
claude-opus-4-8claude-award輸入 · 標準每 1M TokenUSD 1.25
claude-opus-4-8claude-first-topup輸入 · 標準每 1M TokenUSD 0.15
claude-opus-4-8claude-premium輸入 · 標準每 1M TokenUSD 2.5
claude-opus-4-8claude-stable輸入 · 標準每 1M TokenUSD 1.575
claude-opus-5claude-award輸入 · 標準每 1M TokenUSD 1.25
claude-opus-5claude-first-topup輸入 · 標準每 1M TokenUSD 0.15
claude-opus-5claude-premium輸入 · 標準每 1M TokenUSD 2.5
claude-opus-5claude-stable輸入 · 標準每 1M TokenUSD 1.575
claude-opus-5-5claude-award輸入 · 標準每 1M TokenUSD 1
claude-opus-5-5claude-first-topup輸入 · 標準每 1M TokenUSD 0.12
claude-opus-5-5claude-premium輸入 · 標準每 1M TokenUSD 2
claude-opus-5-5claude-stable輸入 · 標準每 1M TokenUSD 1.26
claude-sonnet-5claude-award輸入 · 標準每 1M TokenUSD 0.5
claude-sonnet-5claude-first-topup輸入 · 標準每 1M TokenUSD 0.06
claude-sonnet-5claude-premium輸入 · 標準每 1M TokenUSD 1
claude-sonnet-5claude-stable輸入 · 標準每 1M TokenUSD 0.63
claude-sonnet-5-5claude-award輸入 · 標準每 1M TokenUSD 0.5
claude-sonnet-5-5claude-first-topup輸入 · 標準每 1M TokenUSD 0.06
claude-sonnet-5-5claude-premium輸入 · 標準每 1M TokenUSD 1
claude-sonnet-5-5claude-stable輸入 · 標準每 1M TokenUSD 0.63
codex-auto-reviewopenai-award輸入 · 標準每 1M TokenUSD 0.075
codex-auto-reviewopenai-enterprise輸入 · 標準每 1M TokenUSD 1.25
codex-auto-reviewopenai-first-topup輸入 · 標準每 1M TokenUSD 0.0375
codex-auto-reviewopenai-premium輸入 · 標準每 1M TokenUSD 0.325
codex-auto-reviewopenai-stable輸入 · 標準每 1M TokenUSD 0.2
deepseek-v4-flashdeepseek-award輸入 · 標準每 1M TokenUSD 0.10542168674698796
deepseek-v4-flashdeepseek-stable輸入 · 標準每 1M TokenUSD 0.18072289156626506
deepseek-v4-prodeepseek-award輸入 · 標準每 1M TokenUSD 0.47439759036144574
deepseek-v4-prodeepseek-stable輸入 · 標準每 1M TokenUSD 0.8132530120481928
deepseek-v4.1-flashdeepseek-award輸入 · 標準每 1M TokenUSD 0.10542168674698796
deepseek-v4.1-flashdeepseek-stable輸入 · 標準每 1M TokenUSD 0.18072289156626506
gemini-2.5-flash-imageMedia價格 · 標準每次請求USD 0.032955000000000005
gemini-3-pro-imageMedia輸入 · 標準每 1M TokenUSD 1.6900000000000002
gemini-3.1-flash-imageMedia輸入 · 標準每 1M TokenUSD 0.42250000000000004
gemini-3.1-flash-litegemini-stable輸入 · 標準每 1M TokenUSD 0.0375
gemini-3.1-flash-lite-imageMedia輸入 · 標準每 1M TokenUSD 0.21125000000000002
gemini-3.1-pro-previewgemini-stable輸入 · Context length <= 200000每 1M TokenUSD 0.3
gemini-3.5-flashgemini-stable輸入 · 標準每 1M TokenUSD 0.22499999999999998
gemini-3.5-flash-litegemini-stable輸入 · 標準每 1M TokenUSD 0.045
gemini-3.6-flashgemini-stable輸入 · 標準每 1M TokenUSD 0.22499999999999998
gemini-3.7-flashgemini-stable輸入 · 標準每 1M TokenUSD 0.22499999999999998
gemini-3.8-flashgemini-stable輸入 · 標準每 1M TokenUSD 0.22499999999999998
gemini-3.8-flash-highgemini-stable輸入 · 標準每 1M TokenUSD 0.22499999999999998
gemini-4-argongemini-stable輸入 · 標準每 1M TokenUSD 0.6
glm-5.2glm-award輸入 · 標準每 1M TokenUSD 0.42168674698795183
glm-5.2glm-stable輸入 · 標準每 1M TokenUSD 0.6024096385542169
glm-5.3glm-award輸入 · 標準每 1M TokenUSD 0.42168674698795183
glm-5.3glm-stable輸入 · 標準每 1M TokenUSD 0.6024096385542169
glm-5.3-flashglm-award輸入 · 標準每 1M TokenUSD 0.042168674698795185
glm-5.3-flashglm-stable輸入 · 標準每 1M TokenUSD 0.06024096385542169
gpt-5.5openai-award輸入 · 標準每 1M TokenUSD 0.15
gpt-5.5openai-enterprise輸入 · 標準每 1M TokenUSD 2.5
gpt-5.5openai-first-topup輸入 · 標準每 1M TokenUSD 0.075
gpt-5.5openai-premium輸入 · 標準每 1M TokenUSD 0.65
gpt-5.5openai-stable輸入 · 標準每 1M TokenUSD 0.4
gpt-5.6-solopenai-award輸入 · Context length <= 272000每 1M TokenUSD 0.15
gpt-5.6-solopenai-enterprise輸入 · Context length <= 272000每 1M TokenUSD 2.5
gpt-5.6-solopenai-first-topup輸入 · Context length <= 272000每 1M TokenUSD 0.075
gpt-5.6-solopenai-premium輸入 · Context length <= 272000每 1M TokenUSD 0.65
gpt-5.6-solopenai-stable輸入 · Context length <= 272000每 1M TokenUSD 0.4
gpt-5.6-terraopenai-award輸入 · Context length <= 272000每 1M TokenUSD 0.06
gpt-5.6-terraopenai-enterprise輸入 · Context length <= 272000每 1M TokenUSD 1
gpt-5.6-terraopenai-first-topup輸入 · Context length <= 272000每 1M TokenUSD 0.03
gpt-5.6-terraopenai-premium輸入 · Context length <= 272000每 1M TokenUSD 0.26
gpt-5.6-terraopenai-stable輸入 · Context length <= 272000每 1M TokenUSD 0.16
gpt-6-astraopenai-award輸入 · Context length <= 272000每 1M TokenUSD 0.3
gpt-6-astraopenai-enterprise輸入 · Context length <= 272000每 1M TokenUSD 5
gpt-6-astraopenai-first-topup輸入 · Context length <= 272000每 1M TokenUSD 0.15
gpt-6-astraopenai-premium輸入 · Context length <= 272000每 1M TokenUSD 1.3
gpt-6-astraopenai-stable輸入 · Context length <= 272000每 1M TokenUSD 0.8
gpt-6-lunaopenai-award輸入 · Context length <= 272000每 1M TokenUSD 0.0075
gpt-6-lunaopenai-enterprise輸入 · Context length <= 272000每 1M TokenUSD 0.125
gpt-6-lunaopenai-first-topup輸入 · Context length <= 272000每 1M TokenUSD 0.00375
gpt-6-lunaopenai-premium輸入 · Context length <= 272000每 1M TokenUSD 0.0325
gpt-6-lunaopenai-stable輸入 · Context length <= 272000每 1M TokenUSD 0.02
gpt-6-solopenai-award輸入 · Context length <= 272000每 1M TokenUSD 0.06
gpt-6-solopenai-enterprise輸入 · Context length <= 272000每 1M TokenUSD 1
gpt-6-solopenai-first-topup輸入 · Context length <= 272000每 1M TokenUSD 0.03
gpt-6-solopenai-premium輸入 · Context length <= 272000每 1M TokenUSD 0.26
gpt-6-solopenai-stable輸入 · Context length <= 272000每 1M TokenUSD 0.16
gpt-6.1-solopenai-award輸入 · Context length <= 272000每 1M TokenUSD 0.06
gpt-6.1-solopenai-enterprise輸入 · Context length <= 272000每 1M TokenUSD 1
gpt-6.1-solopenai-first-topup輸入 · Context length <= 272000每 1M TokenUSD 0.03
gpt-6.1-solopenai-premium輸入 · Context length <= 272000每 1M TokenUSD 0.26
gpt-6.1-solopenai-stable輸入 · Context length <= 272000每 1M TokenUSD 0.16
gpt-6.1-terraopenai-award輸入 · Context length <= 272000每 1M TokenUSD 0.03
gpt-6.1-terraopenai-enterprise輸入 · Context length <= 272000每 1M TokenUSD 0.5
gpt-6.1-terraopenai-first-topup輸入 · Context length <= 272000每 1M TokenUSD 0.015
gpt-6.1-terraopenai-premium輸入 · Context length <= 272000每 1M TokenUSD 0.13
gpt-6.1-terraopenai-stable輸入 · Context length <= 272000每 1M TokenUSD 0.08
gpt-image-2Media輸入 · 標準每 1M TokenUSD 3.77
gpt-image-2.5-flareMedia輸入 · 標準每 1M TokenUSD 5.33
gpt-image-2.5-sunburstMedia輸入 · 標準每 1M TokenUSD 5.33
grok-4.5grok-award輸入 · 標準每 1M TokenUSD 0.02
grok-4.5grok-stable輸入 · 標準每 1M TokenUSD 0.22
grok-4.6grok-award輸入 · 標準每 1M TokenUSD 0.02
grok-4.6grok-stable輸入 · 標準每 1M TokenUSD 0.22
grok-4.7grok-award輸入 · 標準每 1M TokenUSD 0.02
grok-4.7grok-stable輸入 · 標準每 1M TokenUSD 0.22
happyhorse-1.1-i2vMedia輸出 · 1080p每秒USD 0.1287
happyhorse-1.1-r2vMedia輸出 · 1080p每秒USD 0.1287
happyhorse-1.1-t2vMedia輸出 · 1080p每秒USD 0.1287
kimi-k3kimi-award輸入 · 標準每 1M TokenUSD 1.0542168674698795
kimi-k3kimi-stable輸入 · 標準每 1M TokenUSD 1.957831325301205
kling-2.5-turboMedia輸出 · 1080p每秒USD 0.0637
kling-2.6Media輸出 · 1080p每秒USD 0.0637
kling-3.0Media輸出 · 1080p每秒USD 0.10192
kling-3.0-omniMedia輸出 · 1080p每秒USD 0.10192
kling-3.0-turboMedia輸出 · 1080p每秒USD 0.1274
kling-o1Media輸出 · 1080p每秒USD 0.10192
qwen3.8-maxqwen-stable輸入 · 標準每 1M TokenUSD 1.4457831325301207
seedance-2.0Media輸出 · 1080p每 1M TokenUSD 9.509500000000001
seedance-2.0-fastMedia輸出 · 480p每 1M TokenUSD 3.435714
seedance-2.0-miniMedia輸出 · 480p每 1M TokenUSD 1.708571
seedance-2.5Media輸出 · 1080p每 1M TokenUSD 14.4495
seedream-5-0-lite-260128Media價格 · 標準每張圖片USD 0.033215
wan2.7-i2vMedia輸出 · 1080p每秒USD 0.11115000000000001
wan2.7-imageMedia價格 · 標準每次請求USD 0.021192600000000002
wan2.7-image-proMedia價格 · 標準每次請求USD 0.0529815
wan2.7-r2vMedia輸出 · 1080p每秒USD 0.11115000000000001
wan2.7-t2vMedia輸出 · 1080p每秒USD 0.11115000000000001
wan3.0-videoMedia輸出 · 1080p每秒USD 0.1976
wan3.0-video-primeMedia輸出 · 1080p每秒USD 0.3094

Pricing and API questions

How do I calculate the cost of a model request?

Choose the model and route, then check the billing unit and conditions. Text models usually list separate USD rates per million input and output tokens. Cache use, context length and request options may change the rate. Usage records show the settled cost of each request.

Why does the same model have several prices?

A model can have several routing groups with different rates and access conditions. Compare the group, supported protocol and any eligibility requirements together. Promotional rates are subject to their stated conditions; your account may have different access or discounts.

How are image and video requests billed?

Billing depends on the model: per image, per second, per request or by tokens. Resolution, quality, duration, audio and reference inputs can affect the charge. Open the model details and API documentation to check the applicable conditions before sending a request.

Do Castaly generations and API requests use the same balance?

Yes. Castaly at castaly.ai uses your Modelflare account and wallet for image and video generation. Modelflare provides direct API access, API keys and usage records. Review the selected model and request settings in either entry point.

What should I check after a failed or timed-out request?

Check the request or task status, usage record and balance changes. A client timeout does not by itself confirm that an upstream task stopped. If the status and charge do not match, contact support with the request or task ID; never include your API key.

Can every model use the same API format?

No. Supported endpoints and parameters depend on the model and route. Check the model documentation for the exact model ID, endpoint, required inputs and streaming or polling behavior. A model listing does not mean every protocol or parameter is supported.

claude-first-topup

Exclusive Claude pricing unlocked when any single top-up reaches US$50. Ideal for personal development, feature validation, and light usage.

openai-first-topup

Exclusive GPT pricing unlocked when any single top-up reaches US$20. Ideal for personal development, feature validation, and light usage.