Bailian applies dynamic rate limiting to some models. The TPM rate limit value is adjusted monthly based on your Bailian monthly consumption tier, and the actual available TPM is no lower than the rate limit value.
Dynamic rate limiting mechanism
Bailian applies a dynamic rate limiting mechanism to some models. The key points are as follows:
- Rate limit value: Determined by tier based on your monthly consumption amount on Bailian. It serves as the guaranteed baseline for throughput and is adjusted every month.
- Soft rate limiting: When the call volume reaches the rate limit value, if the platform still has spare resources, rate limiting is not triggered. Therefore, the actual available TPM is no lower than the rate limit value. The actual rate limit value observed during calls may be higher than this rate limit value, which is expected behavior.
- Rate limit granularity: By account + model. The call volumes of all workspaces and API Keys under a primary account are aggregated and counted toward the TPM quota for that account for that model, sharing the same limit.
- Workspace: You can set a separate rate limit value for a model in a workspace. Workspaces without a separate setting share the account-level quota.
- RPM: Under the dynamic rate limiting mechanism, a model has a relatively large RPM limit. You do not need to set it separately, and RPM rate limiting is not triggered during normal use.
- More information: For the basic concepts of RPM/TPM and details by dimension, see Rate limiting. For workspace rate limit settings, see Permission management.
Models and rate limit tiers
Bailian sets your TPM rate limit tier based on your historical consumption. For the actual rate limit value, refer to the model details page.
- Singapore
- China (Beijing)
Supported models | TPM rate limit tiers | ||
|---|---|---|---|
2,000,000 | 5,000,000 | 10,000,000 | |
qwen3.8-flash | 2,500,000 | 5,000,000 | 10,000,000 |
kimi-k3 | 1,200,000 | 5,000,000 | 10,000,000 |
qwen3.8-27b | 5,000,000 | 5,000,000 | 5,000,000 |
qwen3.8-2.4t-a95b | 5,000,000 | 5,000,000 | 5,000,000 |
1,200,000 | 5,000,000 | 10,000,000 | |
Models and rate limit tiers
The TPM rate limit value is tiered based on the monthly consumption amount on the Bailian platform. The tiers for supported models in different regions are as follows (values are subject to the actual display in the console):
- China (Beijing)
- Singapore
Bailian platform monthly consumption | qwen3.8-max | qwen3.8-flash | qwen3.8-27b | qwen3.8-2.4t-a95b | deepseek-v4-pro-0813 | kimi-k3 |
|---|---|---|---|---|---|---|
≤ 100,000 | 5,000,000 | 5,000,000 | 5,000,000 | 5,000,000 | 1,200,000 | 1,200,000 |
(100,000, 1,000,000] | 10,000,000 | 10,000,000 | 5,000,000 | 5,000,000 | 10,000,000 | 5,000,000 |
(1,000,000, +∞) | 20,000,000 | 20,000,000 | 5,000,000 | 5,000,000 | 20,000,000 | 10,000,000 |
Quota display
Log in to the Bailian console. In the left-side navigation pane, choose Quota management to go to the quota management page.
The quota management page is a read-only page that displays the models that support dynamic rate limiting, the current TPM rate limit value, and the Bailian consumption tier.
Dynamic rate limit adjustment strategy
Rate limit tiers are dynamically adjusted according to the following rules:
- Tiering basis: Bailian platform consumption amount.
- Calculation cycle: Calculated by calendar-month billing cycle.
- Notification: A notification is sent to the affected users on the 10th of each month.
- Effective time: The new tier takes effect on the 15th of each month.
- Adjustment direction: Lower tier → higher tier (upgrade) or remain unchanged.