LIVE FREE TOKEN & PROMO TRACKER•Synced 2026-11-22 13:48 UTC

Free AI API Tokens, Credits & Live Promos

Stop overpaying while prototyping. Browse verified zero-card free API tiers, auto-synced $0/1M OpenRouter models, China-direct free quotas, and up to $350K in official cloud startup grants.

Submit a Free Deal
Verified Channels
16
Official APIs & Grants
No Credit Card Needed
13
Instant API Key Access
Live $0/1M Models
14
Live from OpenRouter API
Max Startup Credits
$350K+
Azure / GCP / AWS Programs
01 / AUTO-SYNCED ZERO-COST ENDPOINTS

Live $0 / 1M Token Models (OpenRouter :free)

Automatically refreshed hourly & daily from OpenRouter's live catalog. Click Copy cURL to test any model in 10 seconds.

Google

Lyria 3 Pro Preview

$0 / 1M
google/lyria-3-pro-preview
Context: 1.0M
Open
Google

Lyria 3 Clip Preview

$0 / 1M
google/lyria-3-clip-preview
Context: 1.0M
Open
Google

Gemma 4 26B A4B (free)

$0 / 1M
google/gemma-4-26b-a4b-it:free
Context: 262K
Open
Google

Gemma 4 31B (free)

$0 / 1M
google/gemma-4-31b-it:free
Context: 262K
Open
NVIDIA

Nemotron 3.5 Super VL (free)

$0 / 1M
nvidia/nemotron-3.5-super-vl:free
Context: 1.0M
Open
NVIDIA

Nemotron 3.5 Lightning (free)

$0 / 1M
nvidia/nemotron-3.5-lightning:free
Context: 1.0M
Open
NVIDIA

Nemotron 3 Ultra (free)

$0 / 1M
nvidia/nemotron-3-ultra-550b-a55b:free
Context: 1.0M
Open
NVIDIA

Nemotron 3 Nano Omni (free)

$0 / 1M
nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free
Context: 256K
Open
inclusionAI

Ling 3.2 Flash (free)

$0 / 1M
inclusionai/ling-3.2-flash:free
Context: 1.0M
Open
02 / VERIFIED PROVIDER PATHS & CREDITS

Free Tiers, CN Direct Quotas, Startup Grants & Discounts

GoogleNo Card

Google AI Studio — Free Gemini API Tier

Free API Tier
🎁Free RPM / RPD across Gemini Flash & Pro
Gemini 2.5 FlashGemini 2.5 ProGemini Flash-Lite

How to claim: Sign in with a Google account in AI Studio and click 'Get API key'—zero billing setup required.

Note: Free tier prompts may be used to improve Google products; upgrade to Pay-as-you-go for zero-data-retention.

https://generativelanguage.googleapis.com/v1beta/openai/
Global (Proxy needed in CN)Claim / Open Portal
OpenRouterNo Card

OpenRouter :free Route — Multi-Model Zero-Cost Pool

Free API Tier
🎁$0 / 1M tokens on 20+ open-weights models
DeepSeek R1/V3 :freeQwen3 :freeLlama 4 :freeGemma 3 :free

How to claim: Create an OpenRouter key and append ':free' to any eligible model ID from the live pool above.

Note: Rate-limited per minute & per day (higher daily cap after adding $10+ account credits).

https://openrouter.ai/api/v1
GroqNo Card

Groq Cloud — Ultra-Fast Free Inference Tier

Free API Tier
🎁Daily free token pool (300–800 t/s speed)
Llama 4 Scout/MaverickQwen3DeepSeek DistillWhisper v3

How to claim: Sign up at Groq Console with GitHub or Google and generate an OpenAI-compatible API key immediately.

Note: Strict tokens-per-minute (TPM) and requests-per-day limits suited for prototypes & low-volume agents.

https://api.groq.com/openai/v1
GitHubNo Card

GitHub Models — Free Playground & Prototyping API

Free API Tier
🎁Free daily requests with Personal Access Token
GPT-4.1 / MiniDeepSeek R1Llama 4Mistral Large

How to claim: Open GitHub Marketplace → Models, generate a fine-grained PAT with `models:read`, and call the Azure/GitHub endpoint.

Note: Rate-limited by GitHub account tier (Free / Pro / Copilot); strictly for prototyping before production.

https://models.github.ai/inference
CerebrasNo Card

Cerebras Inference — 1M Free Tokens / Day

Free API Tier
🎁Up to 1,000,000 tokens/day @ 1,500+ t/s
Llama 3.3 70BQwen3 32B/235BLlama 4

How to claim: Register on Cerebras Cloud with email/GitHub to unlock instant free-tier API keys.

Note: Per-minute token caps apply; ideal for low-latency coding autocomplete & multi-step agent loops.

https://api.cerebras.ai/v1
Mistral AINo Card

Mistral La Plateforme — Experiment Free Plan

Free API Tier
🎁Free API tier + dedicated free Codestral endpoint
Mistral SmallMistral LargeCodestralPixtral

How to claim: Create a Mistral Console workspace, verify phone number, and select the Experiment (Free) plan.

Note: Requires opting into data usage on the Experiment tier; rate limits apply per second.

https://api.mistral.ai/v1
CloudflareNo Card

Cloudflare Workers AI — 10,000 Free Neurons / Day

Free API Tier
🎁10,000 Neurons/day forever free (edge global)
Llama 3.3 / 4Qwen2.5 / Qwen3BGE EmbeddingsWhisper

How to claim: Sign up for a free Cloudflare account and call Workers AI via REST API or Workers binding.

Note: Best for lightweight classification, embeddings, and small-model RAG; heavy models consume Neurons faster.

SiliconFlow (硅基流动)No Card

SiliconFlow — Signup Credits + Permanent Free Open Models

CN Direct
🎁Free signup balance + $0/1M on selected 7B–9B models
Qwen2.5 / Qwen3 (Free)DeepSeek V3/R1GLM-4-9B (Free)

How to claim: Register on SiliconCloud with phone/email to claim starter credits and call free-tagged models immediately.

Note: Free models have RPM/TPM caps; flagship DeepSeek/Qwen-Max models draw from starter or paid balance.

https://api.siliconflow.cn/v1
China Direct + GlobalClaim / Open Portal
Zhipu AI (智谱清言)No Card

Zhipu BigModel — Permanent Free Flash Models & Starter Pack

CN Direct
🎁GLM-4-Flash & CogView/Embedding free tiers
GLM-4-Flash (Free)GLM-4V-FlashGLM-4-Plus

How to claim: Sign in to open.bigmodel.cn, complete account verification, and generate an OpenAI-compatible API key.

Note: Free Flash models enforce concurrency limits suitable for indie apps and automation workflows.

https://open.bigmodel.cn/api/paas/v4/
China Direct + GlobalClaim / Open Portal
Alibaba Cloud (阿里云百炼)No Card

Alibaba Cloud Model Studio — 1M Free Tokens per New Model

CN Direct
🎁1,000,000 free tokens per eligible Qwen/DeepSeek model
Qwen3 / Qwen-MaxDeepSeek R1/V3Wan / Paraformer

How to claim: Activate Alibaba Cloud Bailian (Model Studio) and claim per-model free trial quotas valid for 90–180 days.

Note: Free quota is per-model and time-limited after activation; monitor expiration in console.

https://dashscope.aliyuncs.com/compatible-mode/v1
China Direct + SingaporeClaim / Open Portal
Volcengine (火山方舟)No Card

Volcengine Ark — 500K Free Tokens per Model + Collaboration Plan

CN Direct
🎁500K free tokens per model (Doubao / DeepSeek)
Doubao 1.5 ProDeepSeek R1/V3Doubao Vision

How to claim: Sign up on Volcengine Ark Console, activate model endpoints, and use the built-in free quota per model.

Note: Disable pay-as-you-go overflow or set budget alerts if you strictly want zero-spend usage.

https://ark.cn-beijing.volces.com/api/v3
Microsoft AzureNo Card

Microsoft for Startups Founders Hub — Up to $150K Credits

Startup Grant
🎁$1,000–$150,000 Azure credits usable on Azure OpenAI
GPT-4.1 / o3DeepSeek on AzureGitHub Enterprise

How to claim: Apply online with LinkedIn & a short product description—no prior VC funding required for Tier 1 ($1K–$5K).

Note: Must be building a software product/SaaS; credits expire after 12 months if unused.

Global (Idea-stage welcome)Claim / Open Portal
Google Cloud

Google for Startups Cloud Program — Up to $350K AI Track

Startup Grant
🎁$2,000 (Start) to $350,000 (Funded AI startups)
Vertex AI Gemini 2.5Claude on VertexCloud Run / GPU

How to claim: Submit your company domain & pitch via the Google for Startups portal; approved credits apply to Vertex AI APIs.

Note: Requires a valid company domain and GCP billing profile; highest tier requires verified institutional funding.

Amazon AWS

AWS Activate — Up to $100K Credits for Bedrock (Claude / Llama)

Startup Grant
🎁$1,000 (Founders) to $100,000 (Portfolio) AWS credits
Claude 4 Sonnet/OpusLlama 4Amazon Nova

How to claim: Apply via AWS Activate Founders (self-funded) or with an accelerator Org ID to fund Amazon Bedrock inference.

Note: AWS account requires billing verification; Bedrock third-party model eligibility varies by credit program.

DeepSeekNo Card

DeepSeek API — Off-Peak 50%–75% Discount + Context Caching

Promo / Off-Peak
🎁50%–75% off during UTC 16:30–00:30 + 90% cache savings
DeepSeek-V3 (chat)DeepSeek-R1 (reasoner)

How to claim: Schedule batch summarization, SEO generation, or eval jobs during UTC 16:30–00:30 for automatic off-peak pricing.

Note: Discount windows and percentages are governed by DeepSeek's live pricing page; check before heavy batch runs.

https://api.deepseek.com/v1
China Direct + GlobalClaim / Open Portal
OpenAI & Anthropic

Official Batch API (50% Off) & Prompt Caching (Up to 90% Off)

Promo / Off-Peak
🎁Flat 50% off async jobs + 75%–90% off cached input tokens
Claude 4 SonnetGPT-4.1o3-mini

How to claim: Enable ephemeral prompt caching for long system prompts/docs and route non-realtime workloads to `/v1/batches`.

Note: Batch API completes within 24 hours (not for live chat); caching requires prefix stability.

Spotted a new AI promo, free API tier, or expired credit?

Share it with the community and we'll verify & list it within 24 hours—or subscribe below for weekly price-drop alerts.

Submit Deal / Report Update
UltrAI Weekly Dispatch

Zero hype. Pure AI unit economics & verified money-making playbooks.

Join builders & founders receiving weekly LLM price-drop alerts, A+ audited MCP servers, and MRR teardowns.

1 issue / week. Unsubscribe anytime. Instant access to our Model ROI sheet.