Model Pricing

Showing 3 of 3 models with official list prices. Each entry keeps its own unit and conditions; open a model for every tier.

Published list prices support comparison, while billing uses the approved customer sale-price snapshot for each request.

deepseek-v4-flash

DeepSeek

Output tokens: ¥9.00CNY · per 1M tokens

Conditional price — open the model for details

Input tokens: ¥3.00CNY · per 1M tokens

Conditional price — open the model for details

+4 more price tiers

Context: 1000K

Qwen3.8 Flash

Alibaba

Output tokens: ¥1.35CNY · per 1M tokens

Conditional price — open the model for details

Output tokens: ¥2.70CNY · per 1M tokens

Default

+5 more price tiers

Context: 1000K

Alibaba

Output tokens: ¥18.00CNY · per 1M tokens

Conditional price — open the model for details

Output tokens: ¥36.00CNY · per 1M tokens

Default

+5 more price tiers

Context: 1000K

Model	Developer	Official price tiers	Context
	DeepSeek	Output tokens: ¥9.00CNY · per 1M tokens Conditional price — open the model for details Input tokens: ¥3.00CNY · per 1M tokens Conditional price — open the model for details +4 more price tiers	1000K
	Alibaba	Output tokens: ¥1.35CNY · per 1M tokens Conditional price — open the model for details Output tokens: ¥2.70CNY · per 1M tokens Default +5 more price tiers

Read-only usage quote

Price one explicit request scenario. No upstream call, reservation or charge.

Official list price

ModelMetering modeNative currency (no FX)

Usage per request

Enter actual or expected usage; blank does not mean zero. Token totals and cache components must use normalized protocol semantics.

Total input tokens *Total output tokens *Cache-read tokensCache-write tokens

Identical requests (1–100,000)

Inference modeRate period

Contract conditions & metering details

Conditions must match the model contract. Set region, inference_mode, resolution or rate_period as needed. Audio requires operation: input or output.

Condition selectors (JSON)Modality & cache details (JSON)

Token details accept prompt_tokens_details, completion_tokens_details, cache_read_details, cache_write_details and cache_creation. Enter each successful image size; exclude failed images.

Quote result

Enter a scenario to calculate. Only server-rated results are shown; missing usage or prices are never guessed.

Uses published list prices, supplied usage and current conditions. Not a payment quote. Excludes sell pricing, key snapshots, promotions, private access, FX and vouchers; does not certify native metering acceptance.

Optional private access service

Core API usage is always pay-as-you-go with no base subscription. A service fee applies only when you explicitly activate BYOK or a private gateway; RouterShift platform fallback remains usage-priced.