Official free tier AI API routers & providers

Nscale is a European AI hyperscaler providing Serverless Inference for large-scale AI deployment. It offers an OpenAI-compatible API to leading open models like Llama and DeepSeek. New users receive a $5 signup credit to start prototyping without initial cost.

models: Llama family, DeepSeek family, gpt-oss-120b

Mistral AI's official platform (La Plateforme) provides access to their open and proprietary models via an OpenAI-compatible API. It features a free 'Experiment' tier for developers and pay-as-you-go pricing for production workloads.

models: Mistral Small 4, Mistral Medium 3.5, Mixtral 8x22B Instruct +3

token.sensenova.cn is the official Token Plan for SenseTime's SenseNova multimodal AI platform. During its public beta, it offers generous free quotas for models like SenseNova 6.8 Flash Lite and U1 Fast, suitable for complex office workflows.

models: SenseNova 6.7 Flash-Lite, SenseNova U1 Fast, DeepSeek V4 Flash +1

The xAI Console is the official developer interface for Elon Musk's Grok models. It offers a usage-based API with prepaid credits, prioritizing performance and direct access to their flagship large language models.

models: grok-beta, grok-2, grok-3 +1

ModelScope (modelscope.cn), backed by Alibaba Cloud, is an open-source model community and inference platform. It provides a free API tier allowing users to make up to 2,000 OpenAI-compatible calls per day across a wide range of open-source models (Qwen, DeepSeek, GLM).

models: Qwen family (most stable), DeepSeek-V3.1, Kimi +2

Groq Cloud provides ultra-fast LLM inference using its proprietary LPU (Language Processing Unit) hardware. It offers an OpenAI-compatible API with a substantial free tier for developers to test and build applications at scale.

models: Qwen family, Whisper Large v3, Llama-3.1-8B-Instruct +1

Pioneer AI, by Fastino Labs, provides an inference API for specialized tasks like data extraction and classification. It offers seat-based subscriptions with included platform credits and priority support.

Baseten is a robust model inference and deployment platform that enables enterprises to run custom and open-weights models in production. It offers high scalability, dedicated GPU infrastructure, and a developer-friendly API for deploying models from a comprehensive library.

models: Any supported model in their library (billed by compute; no fixed free-model list)

Cerebras Inference is powered by wafer-scale WSE chips, offering world-leading inference speeds. It provides an OpenAI-compatible API with $5 in free credits and generous rate limits for developers. Details were checked against the provider's own page at www.cerebras.ai.

models: gpt-oss-120b, Qwen3 32B, Llama 4 Scout +1

Cerebras Inference offers ultra-fast AI inference powered by its Wafer-Scale Engine (WSE-3) technology, achieving record-breaking speeds up to 3000 tokens/s. The official API is OpenAI-compatible and features a generous free tier alongside professional pay-as-you-go access.

models: Llama 3.1/3.3 family, Qwen 3 family (per official model list), Llama 4 Scout +1

Z.ai is Zhipu AI's international developer platform, offering access to the GLM (General Language Model) family. It provides a high-performance alternative to Western models, with the GLM-Flash series offered permanently for free to developers to encourage global adoption of its flagship reasoning models.

models: GLM-4.5-Flash, GLM-4.6V-Flash, GLM-4.7-Flash

Inference.net (formerly Kuzco) is a distributed GPU inference network offering a unified OpenAI-compatible API for open-source models. It specializes in low-cost deployment of Llama 3.1 models and offers substantial initial credits ($1 + $25 for surveys) for new developers.

models: Meta-Llama-3.1-70B-Instruct

xAI (founded by Elon Musk) offers the official API for the Grok series of LLMs. Known for leading benchmarks in coding and reasoning, the Grok-4.6 flagship model is available via an OpenAI-compatible API, featuring low-latency inference and high agentic tool-calling capabilities.

models: Grok 4.20

Alibaba Cloud Model Studio (Bailian) is the official international platform for Qwen and other foundation models. It offers 1 million free tokens per model for new users and a 50% discount on batch inference.

Google AI Studio (ai.google.dev) is the official developer platform for Google's Gemini model family. It offers both a generous free tier for prototyping and a paid production tier with higher rate limits, context caching, and data privacy guarantees.

models: Gemma 4 26B A4B , Gemma 4 31B, Gemini 2.5 Flash Lite +3

Lightning AI (LitAI) provides a unified gateway to frontier models through an OpenAI-compatible API. New users receive 40 million free tokens (approx. $15 in credits) to test models like Claude, GPT, and Gemini. Usage beyond the initial credit is billed as pay-as-you-go.

models: gpt-5 / gpt-5.5 family, gemini-3.5-flash / gemini-2.5-pro, nemotron-3-ultra-550b +3

NLP Cloud is a commercial provider hosting various open-model APIs (Llama, Mixtral, Dolphin) with both pay-as-you-go and subscription options. It grants new users $15 in free credit upon phone verification.

models: Llama 3.1 405B, gpt-oss-120b, Llama 3.3 70B Instruct

StepFun Open Platform is the official developer platform of a Chinese AI unicorn, serving its Step-series multimodal models via an OpenAI-compatible API. It supports text, vision, and end-to-end speech models with tiered rate limits and a credits-based free allowance for new accounts.

models: Step-3, Step-series text LLMs, Step-series multimodal/vision models +1

AI21 Labs is an established LLM vendor known for its Jamba model family. The developer platform (api.ai21.com) offers OpenAI-style endpoints for its high-performance models, with a trial credit for new accounts to explore its capabilities.

models: Jamba Large (1.7), Jamba Mini (1.7), Jamba 1.6 Large

Upstage Console provides an OpenAI-compatible API for their high-performance Solar models. It features a developer-friendly $10 signup credit and commitment-based tiers for businesses requiring higher rate limits and support.

models: Solar Pro, Solar Mini, Solar Pro 3

Inception Labs offers a cloud inference platform for its in-house diffusion-based LLMs, such as Mercury 2 and Mercury Coder. Designed for extreme efficiency, it provides an OpenAI-compatible API that significantly undercuts GPU-based inference costs while maintaining high reasoning quality.

models: mercury (general chat diffusion model), mercury-coder (code, with fim/completions), mercury-2

Ollama Cloud is the official managed inference platform for the Ollama ecosystem. It allows developers to deploy and call popular open-weights models (Llama, DeepSeek, Qwen) via a unified OpenAI-compatible API, providing a bridge between local development and cloud scale.

models: deepseek-v3.1:671b-cloud, gpt-oss-120b, gpt-oss:20b +1

InternLM (Shanghai AI Laboratory) provides an open platform with an OpenAI-compatible API. It offers a monthly free token allowance for its InternVL and long-thinking reasoning models. Details were checked against the provider's own page at internlm.intern-ai.org.cn.

models: intern-latest, intern-s1, intern-s1-mini +2

Modal (modal.com) is a serverless compute platform optimized for AI infrastructure. Rather than a per-token API, it bills by the GPU-second, allowing developers to deploy custom models. A 'Starter' tier provides $30/month in free compute credits.

models: No fixed free-model list (self-deploy any supported model, billed by compute)

Openference API offers a high-performance inference platform with a 3-day free trial for agentic development. It features subscription tiers (Lite/Pro/Max) with bundled requests and on-demand rates for extra usage.

models: glm-5.3-flash, deepseek-v4-flash-0731, DeepSeek V4 Flash Vision Exp +3

SambaNova Cloud provides fast LLM inference on specialized RDU hardware. It offers an OpenAI-compatible API with a persistent free tier (Developer Tier) and pay-as-you-go credits for higher rate limits.

models: Llama 3.1 405B, DeepSeek-V3.1/V3.2, gpt-oss-120b +2

Vercel AI Gateway is a unified API gateway that allows routing to hundreds of models from various providers through a single OpenAI-compatible endpoint. It provides observability, budgets, and zero-data-retention (ZDR) options with no token markup over upstream list prices.

models: Access to OpenAI/Anthropic/Google/Meta/DeepSeek and more via the gateway (at upstream list prices)

Dahl Inference is a high-performance platform offering decentralized access to open models like DeepSeek and MiniMax. It features an OpenAI-compatible API and a generous signup promotion of 100 million free tokens.

models: Llama-3.1-8B-Instruct, Qwen2.5 72B Instruct

Qiniu Cloud AI LLM Inference is a managed MaaS platform that provides a unified, compliant API for over 50 mainstream LLMs. It is compatible with both OpenAI and Anthropic protocols, offering new users substantial free tokens and flexible resource packages for production scaling.

models: DeepSeek-V4 series, Qwen, GLM (Zhipu) +3

SiliconFlow (SiliconCloud) is a premier Chinese AI model platform offering high-performance inference for a wide range of open-source models (Qwen, Llama, DeepSeek). It is notable for providing free, permanent API access to models with parameters under 9B, alongside competitive pay-as-you-go rates for larger flagship

models: Qwen2.5-7B-Instruct, Llama-3.1-8B-Instruct

The official Moonshot AI developer platform for Kimi models. Offers pay-as-you-go access to Kimi K3, K2.7-Code, and K2.6 over an OpenAI-compatible API. One-off trial credits available upon verification.

Official Moonshot AI developer platform for Kimi models. Offers pay-as-you-go access to Kimi K3, K2.7-Code, and K2.6 over an OpenAI-compatible API. One-off trial credits available upon verification. Details were checked against the provider's own page at platform.moonshot.cn.

models: moonshot-v1-8k, moonshot-v1-32k, moonshot-v1-128k +1

Cohere provides enterprise-ready LLMs optimized for search, RAG, and agentic workflows. Its official API offers free Trial keys for development and evaluation, alongside high-performance production endpoints for models like Command R+ and the Aya multilingual series.

models: Command R (08-2024), Command R+ (08-2024), Command R+ +2

GitHub Models (models.github.ai) was an official model-inference platform by GitHub for prototyping. As of July 30, 2026, the service has been fully retired and the catalog/API are no longer available. Users are directed to Azure AI for production workloads.

models: Llama family, GPT-4o, GPT-4o-mini

Bytez is an AI model hosting platform that offers an OpenAI-compatible API for a wide range of models. It provides $1 in free credits that refresh every 4 weeks and uses a low-cost pay-as-you-go pricing model.

models: Llama-3.1-8B-Instruct, Anthropic: Claude 3.5 Sonnet, GPT-4o

Fireworks AI is a high-speed serverless inference platform for open-weights models (Llama, Qwen, DeepSeek). It provides an OpenAI-compatible API optimized for low latency and high throughput. Details were checked against the provider's own page at docs.fireworks.ai.

models: Llama family, DeepSeek family, Qwen family

Hyperbolic (hyperbolic.ai) is an open-access AI cloud providing high-performance serverless inference and GPU rentals. It offers an OpenAI-compatible API to 25+ open-source models including DeepSeek, Qwen, and Llama 3.1 405B. New users receive roughly $1 in trial credits upon phone verification.

models: DeepSeek V3, Llama 3.1 405B (BF16, currently the only public 405B Base), Qwen family +2

Hyperbolic (hyperbolic.xyz) is the API-specific endpoint for Hyperbolic's AI cloud. It provides serverless inference for over 25 open-source models through an OpenAI-compatible /v1 endpoint. The platform supports text, image, and audio generation with pay-as-you-go pricing.

models: DeepSeek V3, Llama 3.1 405B, Qwen family +1

Nebius AI Studio is a European AI cloud offering an OpenAI-compatible API to 60+ open-source LLMs; it provides high-performance inference-as-a-service with European data residency and transparent pay-as-you-go pricing.

models: DeepSeek family, Llama/Qwen and 60+ open-source models, Qwen3-235B-A22B +1

Nebius AI Studio is a European-based GPU cloud platform providing managed inference for top open-source LLMs like Llama 3.1 and DeepSeek R1. It leverages its own EU data centers to offer high-performance, privacy-compliant inference with transparent pay-as-you-go pricing tailored for both developers and enterprises.

models: DeepSeek family, Qwen family, Llama-3.1-8B-Instruct

OVHcloud AI Endpoints is an inference API from the established European cloud provider OVHcloud. It offers OpenAI-compatible access to open-weight models with anonymous and registered free tiers for developers.

models: Qwen, Mistral, Llama +1

iFlytek Spark (Xunfei Xinghuo) is a large model platform providing access to the Spark LLM family. It offers a permanently free Spark Lite version and significant one-time token allowances for newer users. The API is OpenAI-compatible and supports various modalities including text, image, and voice.

models: Spark Lite (permanently free), Spark Max (limited/campaign free), Spark Pro/other versions (limited free tokens)

DeepSeek is a leading AI research lab providing frontier-class models (V4-Flash/Pro, R1) via an OpenAI-compatible API. It is known for industry-leading cost efficiency and specialized thinking/reasoning modes.

models: DeepSeek Chat, DeepSeek-R1, DeepSeek V4 Flash Vision Exp +2

Novita AI is a developer-focused inference cloud offering an OpenAI-compatible API to models like DeepSeek, Qwen, and Llama. New users receive a $0.50 trial credit; service is otherwise pay-as-you-go with very low latency.

models: DeepSeek family, Qwen family, Llama family

Scaleway Generative APIs offer a serverless, OpenAI-compatible endpoint for deploying open LLMs. Hosted in Europe, it emphasizes data sovereignty and grants 1,000,000 free tokens to all customers to jumpstart development.

models: Mistral family, Qwen family, Pixtral 12B (2409) +3

Cloudflare Workers AI allows running AI models on Cloudflare's global network. It provides an OpenAI-compatible API for inference, embeddings, and image generation, billed in compute 'Neurons'. Details were checked against the provider's own page at developers.cloudflare.com.

models: Gemma family, Llama family, Qwen family +3

Hugging Face is the leading hub for open-source AI, offering model hosting, datasets, and serverless inference. It provides a permanent free tier for basic CPU spaces and community GPU grants, with professional-grade Inference Endpoints available on a pay-as-you-go basis.

models: Llama family, Qwen family, Gemma family +2

Nomic AI provides the Atlas Embedding API and open-weight models like nomic-embed. It offers a monthly free tier for its embedding API and is transitioning to focus on document AI for specialized industries.

models: nomic-embed-text-v1.5, nomic-embed-text, nomic-embed-vision

Api.airforce is a unified AI model aggregator reselling access to 100+ models from Anthropic, OpenAI, and xAI. It offers a large selection of free models (including Grok-3 and Claude 3.7) alongside paid subscription tiers for higher throughput.

models: Gemini 2.5 Flash, DeepSeek Chat, Grok 4.20 +3

Black Forest Labs (api.bfl.ai) provides official API access to the FLUX diffusion model family. It is a credit-based image generation API where users purchase balance to generate high-fidelity images using FLUX.2 and FLUX.3 models.

BFL ML (api.bfl.ml) is the technical endpoint for Black Forest Labs' FLUX model API. It shares the same credit-based billing as the main bfl.ai domain, providing high-performance access to diffusion models for production workloads.

OpenAI Codex is the specialized model family for coding tasks, now integrated into the flagship GPT models. It provides a specialized developer tier for high-performance code generation, planning, and task automation.

models: gpt-5.6-sol

FreeInference.org is a research-focused AI API gateway built at Harvard SEAS MadSys Lab. It provides free OpenAI-compatible access to frontier open models for the research and education community. No credit card is required for access.

models: Qwen3 Coder, MiniMax-M3, DeepSeek V4 Flash +1

FriendliAI is a high-performance AI inference platform specializing in serving frontier open-weight models like GLM, DeepSeek, and Llama with industry-leading speed. It offers serverless Model APIs, dedicated GPU endpoints, and on-premise containers.

models: Meta-Llama-3.1-70B-Instruct, Llama-3.1-8B-Instruct, Mistral Nemo +1

BigModel.cn (智谱AI) is the official platform for GLM series models, offering high-performance text and multimodal (image/video) inference. It features flagship models like GLM-5.3 and provides a permanent free tier for Flash models along with significant token grants for new users.

models: glm-5.3-flash, glm-5.3, GLM 5 Turbo

Morph (morphllm.com) is an AI provider focusing on agentic coding models. Its API offers high-speed code editing and search via proprietary models (Compact, Reflexes) and optimized frontier models (GLM, Kimi), with a limited free tier of 200 requests per month.

models: deepseek-v4-flash-0731, Qwen3.5-122B-A10B, glm-5.3 +3

NVIDIA NIM (Build) is NVIDIA's official hosted model-inference service, providing API access to over 100 frontier open-source models (Llama 3.3, Nemotron, DeepSeek). It features a permanent free tier with a 40 RPM limit, authenticated via a standard API key and accessible through an OpenAI-compatible endpoint.

models: Nemotron 3 Nano 30B A3B, Nemotron 3.5 Lightning, Nemotron 3 Super +3

OpenVecta provides an AI model API with an OpenAI-compatible endpoint. It focuses on simplicity and ease of use, with reports of a free tier or trial credits available for new users. Details were checked against the provider's own page at openvecta.com.

models: Gemma 4 31B, GLM-4.7-Flash, DeepSeek V4 Flash +3

Perplexity provides the Sonar API for usage-based AI search and reasoning. It offers a limited free tier for end-users and a usage-based API for developers, optimized for real-time grounded answers. Details were checked against the provider's own page at perplexity.ai.

models: Perplexity default free search model (basic Sonar class)

OpenTyphoon is a Thai-focused AI model platform by SCB 10X. It provides free access to models like Typhoon 2.5 via its hosted API at opentyphoon.ai. Production-grade access is offered through partners like Together AI and Float16.

Xiaomi MiMo (mimo.mi.com) is the official AI open platform for Xiaomi's in-house large models. It offers an OpenAI-compatible API for the MiMo-V2.5 series, featuring a high-concurrency token plan (¥60 for 50B tokens) and competitive per-million token pricing for Pro and Flash variants.

models: MiMo-V2.5, MiMo-V2-Flash, MiMo-V2-Pro (time-limited/quota-limited free) +1

spark-api-open.xf-yun.com is the OpenAI-compatible HTTP endpoint for iFlyTek's Spark (讯飞星火) LLM. It provides access to multimodal models via the iFlyTek Open Platform. A permanently free 'Lite' version is available for developers.

models: Spark Lite (model=lite), Spark Pro/Max/4.0 Ultra (one-off free token packs only; paid after depletion)