Model APIs Updates

New Features

  • Rate Limits Now Applied to Image, Video, and Audio APIs

    Rate limiting has been enabled for the Image, Video, and Audio API products to ensure reliable performance and fair usage:

    1. IPM (Images Per Minute): Caps the number of images that can be generated per minute.
    2. RPM (Requests Per Minute): Limits total API calls per minute across all endpoints.

    If your requests exceed the set thresholds, rate limit errors will be returned.

    You can check your current limits and request higher quotas directly in the console. Look for the “Increase Limit” button next to each metric if your use case requires more capacity.

    Updated time: April 30, 2025

  • Five Qwen3 Series Models Now Available

    We’ve added five new models from the Qwen3 series on Novita.:

    1. Qwen3 0.6B
    2. Qwen3 1.7B
    3. Qwen3 4B
    4. Qwen3 8B
    5. Qwen3 14B

    Updated time: April 30, 2025

  • New Model: DeepSeek Prover V2 671B

    Novita supports DeepSeek Prover V2 671B via Serverless Endpoints.

    Updated time: April 30, 2025

  • Updated LLM API Rate Limits

    User accounts are now automatically tiered into L1-L5 service levels based on recent top-up amount, with each tier granting different RPM (Requests Per Minute) and TPM (Tokens Per Minute) limits. For more details, please refer to Rate limits - Documentation.

    Updated time: April 30, 2025

  • First Batch of Large-Parameter Qwen3 Models Now Available

    Three large-parameter Qwen3 models have been newly launched:

    1. Qwen3-235B-A22B
    2. Qwen3-30B-A3B
    3. Qwen3-32B

    Updated time: April 29, 2025