Your saved
resources.

GPT-OSS 120B: 200k tokens/day

Fast hosted inference for language and speech models, with a recurring Free plan.

The catch. GPT-OSS 120B also has limits of 30 requests/minute and 8,000 tokens/minute. Limits differ by model and apply at organization level. Upgrading to Developer enables paid usage.

50 requests/day across free models

One API for trying multiple free language models.

The catch. 20 requests/minute. Failed attempts count toward the daily limit. Paid models require credits; free availability is not a production SLA. Check the selected model's data policies.

10,000 Neurons/day

Hosted AI inference, including FLUX.1 schnell image generation, on a shared daily allowance.

The catch. Neurons measure compute, not tokens or images. The allowance is shared across models. Free-plan requests stop at the limit; paid plans charge overages. Some models require paid billing.

10,000 credits/month

A monthly allowance for experimenting with voice generation and audio tools.

The catch. No commercial license on Free. Published free-plan output requires attribution. Credit usage varies by model and feature; unused quota is not a promise of a fixed number of audio minutes.

2,000 code completions/month

AI code completions and limited chat assistance for individual developers.

The catch. Chat and agent features have a limited AI credit allowance. Free uses automatic model selection. This is an editor assistant, not a general-purpose inference API.

Model-specific free quotas

Free-tier access to selected Gemini models through Google AI Studio.

The catch. Not every model has a free tier. Check your project's live limits. Free-tier data may be used to improve Google's products. Image-generation models should not be assumed free because chat access is free.

$30 compute credit/month

Serverless CPU and GPU compute credits for independent developers.

The catch. Card requirement was not established. Starter is pay-as-you-go beyond included compute. Shared Endpoints are excluded and billed from the first request. Configure a workspace budget before running workloads.

$5 promotional credit

One-time promotional inference credit for trying the Cerebras API.

The catch. API and Playground access pause when the credit expires or runs out unless you separately purchase PayGo credits. The offer does not automatically enroll you in paid usage.