MeetNewAI / Free AI API

Free AI API

Model API quotas you can apply for yourself. Standing free tiers reset daily or monthly. Signup grants end when the credit ends.

Checked 2026-10-16. Quotas follow each vendor console.

29 offers
Type
Region
Gate
Model API quotas you can apply for
VendorQuotaTypeRegionGateApply
Google AI Studio

Gemini Flash, Flash-Lite, and Gemma 4. Flash is about 15 requests/minute and 1,500/day.Free-tier content may be used to improve the product.

Standing freeGlobalEmailApply
Groq

gpt-oss 120B / 20B and Qwen3.8 27B. About 30 requests/minute and 1,000/day.

Standing freeGlobalEmailApply
NVIDIA NIM

More than 100 models, including GLM and Nemotron. About 40 requests/minute.Claim it after joining the NVIDIA Developer Program.

Standing freeGlobalPhoneApply
OpenRouter

Models with :free. About 20 requests/minute and 50/day.After the account is funded with $10, free models can reach 1,000 requests a day.

Standing freeGlobalEmailApply
Cloudflare Workers AI

10,000 Neurons a day, shared across Llama, Gemma, GLM, gpt-oss, and others.

Standing freeGlobalEmailApply
Mistral

The free tier is on by default. Public lists record about $10 of credit per month.Content in free mode may be used for training. You can turn that off in the account.

Standing freeGlobalEmailApply
Ollama Cloud

DeepSeek V4, Kimi, MiniMax, and gpt-oss. Quotas are per session and per week.

Standing freeGlobalEmailApply
Cohere

Trial Key. About 20 requests/minute and 1,000/month.Non-commercial use only.

Standing freeGlobalEmailApply
SambaNova

DeepSeek and MiniMax. About 20 requests/minute and 20/day.

Standing freeGlobalEmailApply
LLM7.io

Turbo models can be called with no key. About 60 requests/minute after you claim a free token.

Standing freeGlobalNo signupApply
Kilo Code

A free pool on the gateway. About 200 requests/hour.Requests may be logged upstream.

Standing freeGlobalEmailApply
OpenCode Zen

Free models inside the coding client, including DeepSeek V4 Flash. The list changes.

Standing freeGlobalEmailApply
OVHcloud AI Endpoints

Anonymous access at 2 requests/minute. Qwen, Llama, and Mistral, in European data centers.

Standing freeGlobalNo signupApply
Aion Labs

Character and narrative models. About 15 requests/minute and 20,000 tokens a day.

Standing freeGlobalEmailApply
智谱 GLM

GLM-4.7-Flash and GLM-4.6V-Flash stay free, at concurrency 1. Sign-up also includes a trial pack of about 20 million tokens.The trial pack usually lasts 1 month.

Standing freeChinaIdentity checkApply
魔搭 ModelScope

Free inference after sign-up, about 2,000 requests/day. An Alibaba Cloud account must be linked.

Standing freeChinaIdentity checkApply
硅基流动

Models at 9B and below stay free. Identity verification also includes a 16 yuan voucher.The voucher offer runs through 2026-12-31.

Standing freeChinaIdentity checkApply
讯飞星火

The Lite API stays free, at a low QPS.

Standing freeChinaIdentity checkApply
快手万擎

KAT-Coder-Air stays free. New Pro users get about 20 million tokens.

Standing freeChinaEmailApply
火山方舟

About 500,000 tokens per model. Campus verification can reach about 100 million.

Signup grantChinaIdentity checkApply
阿里云百炼

About 1 million tokens per model, for about 90 days.

Signup grantChinaIdentity checkApply
百度千帆

About 1 million tokens per model, for about 3 months.

Signup grantChinaIdentity checkApply
腾讯混元

A trial pack on first activation, for about 1 year.

Signup grantChinaIdentity checkApply
DeepSeek

About 5 million tokens for new users, for 30 days.

Signup grantChinaEmailApply
Kimi

About 15 yuan in trial credit after verification.

Signup grantChinaIdentity checkApply
美团 LongCat

About 10 million tokens for 30 days after identity verification.

Signup grantChinaIdentity checkApply
MiniMax

A voucher after identity verification, for about 90 days.

Signup grantChinaIdentity checkApply
天翼云息壤

A trial pack of about 25 million tokens, for 1–3 months.

Signup grantChinaIdentity checkApply
PPIO

A voucher of about 5 yuan for new users.

Signup grantChinaEmailApply