AI21 Labs (Jamba)
Innovative SSM-Transformer hybrid models. $10 free developer credits upon sign-up with 256K context.
RESOURCE RADAR
Innovative SSM-Transformer hybrid models. $10 free developer credits upon sign-up with 256K context.
Alibaba Cloud enterprise platform offering Qwen 2.5, Qwen-Max, and Wanx multimodal models with free credits.
AnyAPI.ai: OpenAI-compatible gateway to 400+ models; see free_quota (radar intake).
AssemblyAI: developer API platform; see free_quota (radar intake, details to verify).
Atlascloud: discovered by radar; free tier and models unverified.
Leading Chinese LLM founded by Sogou CEO. 10M tokens on sign-up, specialized in Chinese NLP and healthcare.
Baidu AI Cloud platform. ERNIE-Speed and ERNIE-Lite models are permanently free for all developers.
Baseten: inference / Model APIs platform; see free_quota (radar intake).
World's fastest wafer-scale inference engine (1800+ tokens/s). 1,000,000 free tokens daily.
Decentralized serverless AI compute. Provides instant drop-in open LLM APIs and free exploration tier.
Serverless GPU inference at the edge. 10,000 free Neurons daily (~100K-300K tokens).
Top enterprise NLP provider. Trial Key is permanently free for Command R+ and world-class Rerank models.
Context.dev: web data / scrape API for AI agents; see free_quota (radar intake).
Deepgram: developer API platform; see free_quota (radar intake, details to verify).
Ultra-fast serverless AI inference platform. Free trial credits for new accounts with full OpenAI SDK compatibility.
Global efficiency frontier. 5,000,000 free tokens upon registration for DeepSeek V3 and R1 reasoning models.
ElevenLabs: developer API platform; see free_quota (radar intake, details to verify).
Exa: developer API platform; see free_quota (radar intake, details to verify).
fal.ai: developer API platform; see free_quota (radar intake, details to verify).
Serverless inference cloud with 1,000+ open-source LLMs. Free trial credits included with drop-in OpenAI protocol compatibility.
Firecrawl: discovered by radar; free tier and models unverified.
High-performance inference platform. $1 free credit (~5,000,000 tokens) on sign-up with ultra-low latency.
getimg.ai: developer API platform; see free_quota (radar intake, details to verify).
Official GitHub & Azure playground. Free daily rate limits across GPT-4o, Claude 3.5, and DeepSeek R1.
Gladia: discovered by radar; free tier and models unverified.
Massive compute grant: 2,000,000,000 free tokens for MiniMax-Text-01 with 245K context window.
Official Google API. 1,500 free requests per day (~10M tokens) for Gemini 2.5 Flash with 1M context.
World-record LPU inference (500-800 t/s). Generous permanent free tier with 30 RPM.
Official Serverless Inference API by Hugging Face. Free access to 300+ open-source models with 300 req/hr tier.
Decentralized GPU cloud. $10 free developer credits for DeepSeek R1 671B full-parameter inference.
Inception Labs: Mercury diffusion LLM API; see free_quota (radar intake).
Tsinghua spin-off multi-chip AI platform. Tens of RMB in free credits with domestic GPU acceleration.
Infron: multi-model inference gateway with :free models; see free_quota (radar intake).
Developer toolkit for RAG & LLM web reading. 10,000,000 free tokens for r.jina.ai and s.jina.ai.
Lightning AI: Studios, compute, and Model APIs; see free_quota (radar intake).
Founded by Kai-Fu Lee. 20 RMB (~10M tokens) on registration for Yi-Lightning & Yi-Large.
Linkup: developer API platform; see free_quota (radar intake, details to verify).
LLM7.io: developer API platform; see free_quota (radar intake, details to verify).
Leading autonomous AI Agent platform. Official 1 Trillion free AI tokens campaign for developers to build and host full-stack intelligent web apps.
Trillion-parameter MoE foundation model. 15,000,000 free tokens for MiniMax-Text-01 and TTS.
Europe's leading AI lab. Permanent free experimentation tier with 1 RPS for Mistral Large and Codestral.
Modal: developer API platform; see free_quota (radar intake, details to verify).
Alibaba's open-source model hub. Free Serverless Inference API for Qwen 2.5, DeepSeek, and Llama 3.
Pioneered 2M-token long context in China. 15 RMB voucher (~1.25M tokens) on registration.
European AI cloud infrastructure. $5-$10 free credits upon sign-up for DeepSeek R1 and Llama 3.3.
Fast and affordable serverless LLM platform. $0.5-$2 free credits on sign-up with low pricing.
Official NVIDIA enterprise inference platform. 1,000 free API credits (~40M tokens) hosted on H100 clusters.
Global LLM unified gateway. Permanent free access to 30+ open-source models with ':free' suffix.
Completely free, open-source, keyless multimodal generation API for images and text.
PPIO (ppinfra): China distributed compute and model APIs; see free_quota (radar intake).
Requesty: discovered by radar; free tier and models unverified.
SN40L custom silicon cloud. Permanent free tier for full-parameter DeepSeek R1 671B and Llama 3.3.
Scaleway Generative APIs: EU-hosted LLM/audio APIs; see free_quota (radar intake).
China's national supercomputing AI grid. 10,000,000 free tokens voucher upon developer registration.
Curated Sepolia testnet ETH faucets. Free 0.05-0.5 testnet ETH daily for Web3 developers.
Serper: developer API platform; see free_quota (radar intake, details to verify).
Leading Chinese inference accelerator. 20,000,000 tokens on sign-up plus permanent free models.
Speechmatics: developer API platform; see free_quota (radar intake, details to verify).
Stability AI: developer API platform; see free_quota (radar intake, details to verify).
Leading Chinese Foundation Model provider. 15 RMB free voucher on registration for Step-2 series models.
Search engine optimized for AI agents & LLMs. 1,000 free search API calls every month.
Tencent's foundation model service. Free credits for new accounts with strong multi-turn WeChat reasoning.
SmarToken: OpenAI-compatible gateway for Chinese frontier models; see free_quota (radar intake).
Leading open-source cloud. $5 free trial credits (~25M tokens) for LLaMA, DeepSeek, and FLUX.
High-throughput cloud powered by H200 and Ascend clusters. 2M-5M tokens on sign-up with 1024K context.
Vercel AI Gateway: multi-model gateway; see free_quota (radar intake).
ByteDance cloud platform. 2,000,000 free tokens daily (60M/month) for Doubao and DeepSeek via Ark program.
Elon Musk's xAI platform. $25 sign-up credit plus up to $150/month free credits via Data Sharing program.
iFLYTEK AI platform. Spark Lite model is permanently free for all developers with full OpenAI compatibility.
You: discovered by radar; free tier and models unverified.
Leading Chinese AI laboratory. GLM-4-Flash is permanently free; new users get 25,000,000 free tokens.
Found an unlisted model platform offering free credits? Submit a GitHub issue to get it swept and indexed!
扫码关注官方公众号,第一时间接收全网大模型免费额度变动、新源上线与 Agent 降本内参。

微信扫一扫 · 极客同行