SiliconFlow Free Credits and Free Models
$1 free credit plus permanently free models on SiliconFlow
What you get from SiliconFlow Free Credits and Free Models for free
Overview
SiliconFlow is a model-inference platform offering an OpenAI-compatible API for open-source and commercial text, image, and audio models. The official pricing page states: "Flexible token pricing, high usage limits, and postpaid billing—plus $1 in free credits to get you started!" The pricing FAQ repeats that "You only pay for what you use, and you can start with $1 in free credits."
Free Tier Quotas & Limits
- $1 in free credits on signup, usable on paid models at posted token rates.
- Free models are available separately. The rate-limit documentation states: "The Rate Limits for free models are fixed, while paid models have different Rate Limit metrics based on account usage tiers."
- New accounts start at usage tier L0: 1,000 RPM and 40,000 TPM. Higher tiers follow monthly spend (L1 1,200 RPM / 60,000 TPM through L5 10,000 RPM / 2,000,000 TPM).
- Monthly spending counts both recharged and gifted amounts, and tiers are recalculated by comparing the previous calendar month against the current month to date.
deepseek-ai/DeepSeek-R1anddeepseek-ai/DeepSeek-V3carry an extra limit of 30 requests per hour for all users, plus 100 requests per day for users who have not completed identity verification.- Rate limits apply per user account, not per API key, and each model has separate limits. Exceeding a limit returns HTTP 429.
How to Get Started
- Create an account at the SiliconFlow console.
- Generate an API key on the API Keys page.
- Point any OpenAI client at the base URL
https://api.siliconflow.com/v1and pass your key. - Set a monthly spending limit in the dashboard if you want a hard cap on paid usage.
Before you claim
- Payment card: Not needed.
- Account: Sign up with the provider.
Key limits
- Free-model rate limits are fixed
- New accounts start at usage tier L0: 1,000 RPM / 40,000 TPM
- DeepSeek-R1/V3 limited to 30 requests/hour (100 requests/day without identity verification)
- Limits are per account, not per API key
- Monthly spending raises your tier and limits
Ollama (MIT) runs on your own machine with no quota.
Runs the small open models SiliconFlow serves free.
Models served free
Limits as the provider publishes them; a dash means none is published.
| Model | Context | RPM | RPD | TPM | Tools | Vision |
|---|---|---|---|---|---|---|
| Qwen/Qwen2-7B-InstructOn the free LLM list; limits fixed but unpublished | 32k | — | — | — | — | — |
| THUDM/glm-4-9b-chatFree LLM list | 32k | — | — | — | — | — |
| mistralai/Mistral-7B-Instruct-v0.2Free LLM list | 32k | — | — | — | — | — |
| Qwen/Qwen2-1.5B-InstructFree LLM list | 32k | — | — | — | — | — |
History
Verified by a person on · last automated check : no change found. Automated findings never change this page until a person confirms them (how we verify).
- Verifiedby a person
- Listedby a person
Full history with screenshots · How we verified this offer · Sitewide changelog
Still works?
Work at cloud.siliconflow.com? Add an aifree.dev badge
<a href="https://aifree.dev/offer/siliconflow-free-inference-credits"><img src="https://aifree.dev/badge/siliconflow-free-inference-credits" alt="SiliconFlow Free Credits and Free Models on aifree.dev" height="20"></a>[](https://aifree.dev/offer/siliconflow-free-inference-credits)Step-by-step guide
Understand it
A few minutes of play-along, no sign-up.
- Why does it miscount the r’s in "strawberry"? · Tokens · 3 min
- Why does the same prompt give different answers? · Temperature and sampling · 3 min
Appears in
More free API credits
Related offers
$200 in credits - More than 20 open-weight models (LLM, TTS/STT, document analysis, image analysis)
Worth $20 - 2,000 free credits per account on first API key generation
30M tokens/month on the free tier (input + output, incl. reasoning)
$200 credit valid for 30 days only
500,000 tokens/month - 80 requests/minute and 100,000 tokens/minute on the free plan