AI API Router Directory

Baidu Qianfan is the official Model-as-a-Service (MaaS) platform of Baidu Intelligent Cloud. It serves the ERNIE (Wenxin Yiyan) series models and other third-party models via an OpenAI-compatible v2 API, offering high-performance inference for enterprise and individual developers.

No public free-tier information found (researched 2026-08-29).

Baseten offers Model APIs, hosted open-weight models behind an OpenAI Chat Completions endpoint at inference.baseten.co with a beta Anthropic Messages endpoint, alongside dedicated GPU deployments and training billed per minute. Model API rates range from $0.10/1M input and $0.50/1M output for GPT OSS 120B to $3.00/1M input and $15.00/1M output for Kimi K3, with cached input billed at roughly a tenth of the input rate. The Basic plan is $0 per month pay-as-you-go, and the pricing FAQ states new accounts come with credits but publishes no amount.

The official pricing FAQ states that new Baseten accounts come with credits to experiment with deployments, but no dollar amount, expiry or conditions are published (checked 2026-09-16).

Open-source Go proxy server (MIT license, router-for-me/CLIProxyAPI) that wraps CLI subscriptions - Google Antigravity/Gemini, OpenAI ChatGPT Codex (GPT series), Anthropic Claude Code, xAI Grok Build, Moonshot Kimi - and exposes them as OpenAI (incl. Responses), Gemini (incl. Interactions), Claude, Codex and Grok compatible API endpoints. OAuth login flows, multi-account round-robin load balancing, streaming/non-streaming/WebSocket, function calling, multimodal input, Management API, and an embeddable Go SDK. Self-hosted; users bring their own upstream accounts/keys. Desktop wrapper: EasyCLIProxyAPI.

Software is free (MIT); access to upstream models depends on user-provided subscriptions/API keys, no credits or prices from the proxy itself.

ChatGPT's free tier is available to everyone and is described on the official pricing page as unlimited text chats with GPT-5.6 Luna subject to abuse guardrails, with limited messages when uploading files, limited and slower image generation, limited voice chats, limited deep research, limited memory and context, limited Codex access and limited ChatGPT Work access on desktop. Paid plans are priced per user per month at $8 (Go), $20 (Plus) and $200 (Pro) as announced by OpenAI on 16 January 2026, with Business and Enterprise offered monthly or annually.

No free credit or USD allowance is stated for ChatGPT on the official pricing page or the ChatGPT Go announcement (checked 2026-09-16); the free tier is a feature-limited plan, not a credit.

Chutes is a decentralized serverless AI inference marketplace on Bittensor (Subnet 64). It provides an OpenAI-compatible API for deploying and scaling open-source models using high-performance VLLM templates, with per-token pricing and optional monthly subscription plans.

Official announcement 2026-02-27: $5-deposit 200-requests/day plan closed to new users, fully retires 2026-03-15; now pay-per-token (or GPU by the second). Bittensor-subsidized cheap models are not a free tier.

Clarifai provides a production-ready AI API for developers, offering an OpenAI-compatible endpoint to run inferences on Clarifai-hosted models from labs like Meta, Google, and Mistral, alongside their own specialized vision and NLP models.

No public free-tier information found in the source reference; treat free-tier availability as unverified rather than confirmed absent.

Conol (conol.ai) — knowledge-base product with background research/notes agents. Web app forwards to OpenAI models via its model picker (e.g. GPT-6 Astra). Unofficial/experimental.

A free tier is available; the exact credit value is not quantified.

DeepInfra is an OpenAI-compatible inference cloud for open-weight models, with per-token rates published on a single pricing page. Examples from that page: DeepSeek-V4-Flash-0731 at $0.06 per 1M input tokens ($0.015 cached) and $0.18 per 1M output; Qwen3.5-9B at $0.10/$0.15; DeepSeek-V3.2 at $0.26/$0.38; Kimi-K3 at $2.85 per 1M input and $14.25 per 1M output; embeddings from $0.005 per 1M input tokens. Models without fixed token rates are billed on execution time, and custom LLM deployments are billed per minute of GPU uptime across A100 ($0.89/GPU-hour) through B300 ($4.89/GPU-hour). There is no free credit: the pricing page states you must add a card or pre-pay before you can use the service.

No free tier stated on official pages (checked 2026-09-16): deepinfra.com/pricing states that you must add a card or pre-pay before using the service, with no starter credit offered.

Duck.ai is DuckDuckGo's free, private AI chat (also surfaced at duckduckgo.com/duckchat). It requires no account and offers chat, voice, image, and search across models including OpenAI GPT (e.g., '5.6 Luna') and Claude.

Free anonymous chat; an 'Upgrade'/Pro option exists but pricing not quantified on fetched pages.

FreeInference (freeinference.org) is an OpenAI- and Anthropic-compatible inference API offering frontier open models free of charge to the research community, with no credit card required at signup. It is presented as built at Harvard SEAS (MadSys Lab) and lists NVIDIA and Harvard SEAS as sponsors. The quota is described only as a "generous quota for research and prototyping"; no numeric limit is published, and the /pricing and /limits paths both return 404. The site carries a banner warning that GLM, Minimax and Kimi will be discontinued on Oct 1, and on Aug 11, 2026 it announced that new account onboarding is paused because the service is at capacity.

No free tier stated on official pages (checked 2026-09-16): freeinference.org offers a free account with a "Generous quota for research and prototyping" but publishes no dollar or token figure, and freeinference.org/pricing returns 404.

Opengateway (by Gitlawb) is a pay-as-you-go inference gateway providing unified OpenAI- and Anthropic-compatible access to models from Xiaomi MiMo, Google, MiniMax, and Qwen. It features smart routing via 'model: auto' and a permanent free tier for Nemotron 3 Ultra.

No public free-tier information found in the source reference; treat free-tier availability as unverified rather than confirmed absent.

HelixMind is an AI API router designed to manage and optimize costs across hundreds of LLMs through a single endpoint. It native speaks the OpenAI API format and features automatic model selection by price and task requirement, ensuring enterprise-grade reliability.

No public free-tier information found in the source reference; treat free-tier availability as unverified rather than confirmed absent.

Hyperbolic (hyperbolic.ai) is an AI cloud that sells on-demand, reserved and private GPU capacity and, separately, a serverless inference API described as OpenAI-compatible at base URL https://api.hyperbolic.xyz/v1 with 25+ open-source models. Its official rate-limit table defines a Basic tier of 60 requests per minute available to a free account with a $0 minimum deposit, and a Pro tier of 600 requests per minute that unlocks automatically once $5 or more is deposited. Serverless inference is pay-as-you-go: text generation runs from $0.10 per 1M tokens for small models, $0.20-$0.40 for 32B-72B models and $0.30-$4.00 for 120B-480B models, with images at a $0.01 base rate per 1024x1024 image at 25 steps and audio at $5.00 per 1M characters.

No free tier stated on official pages (checked 2026-09-16): Hyperbolic documents a free "Basic" tier as 60 requests/minute with a Free account, but no dollar credit or trial allowance is published on the pricing or billing pages.

Inference.net is a unified LLM gateway, observability and fine-tuning platform rather than an image or video vendor. Its pay-as-you-go tier is $0 + usage and includes 1M gateway requests and 1M tracing spans per month, 14-day data retention, one seat and a 30 req/min rate limit. Growth is $250/month with a $50 one-time opening credit, 50M gateway requests and tracing spans per month, unlimited seats and a 250 req/min limit. The catalog lists 60 models from OpenAI, Anthropic, Google and open-source labs billed per token, from Schematron V2 Turbo at $0.03/$0.15 to Claude Opus 5 and GPT-6 Astra at $5.00-$10.00 input and $25.00-$50.00 output per 1M tokens.

No free dollar credit: the $0 pay-as-you-go tier includes 1M gateway requests and 1M tracing spans per month, while the $50 one-time opening credit is listed only on the $250/month Growth plan (checked 2026-09-16).

Kenari is an Indonesian AI gateway providing a unified OpenAI-compatible API to 38+ models from Claude, GPT, and Gemini. It is uniquely tailored for the Indonesian market, accepting local payment methods like QRIS and billing in Rupiah (IDR).

No public free-tier information found in the source reference; treat free-tier availability as unverified rather than confirmed absent.

Kilo Code is an open-source (MIT-licensed) agentic coding assistant that runs in VS Code, JetBrains, a CLI, and cloud agents, now branded at kilo.ai after kilocode.ai redirects there. The platform itself is free for individuals: the pricing page lists the Individual plan at $0 with "Usage is billed separately", Teams at $15/user/month and Enterprise as custom, with a 14-day Enterprise trial. Inference is deliberately provider-agnostic - you can use free models, local models via Ollama or LM Studio, or bring your own Anthropic/OpenAI/Google/Azure/AWS keys at $0/mo with "No hosted credit required" - while hosted inference is either pay-as-you-go through Kilo Gateway or a Kilo Pass subscription starting at $19/mo with up to 50% bonus credits.

No free tier stated on official pages (checked 2026-09-16): Kilo publishes a $0 individual plan and a $0/mo free/BYOK/local inference path, but no dollar credit grant; Kilo Pass bonus credits are tied to a paid subscription from $19/mo.

Kilo Gateway is Kilo's OpenAI-compatible inference gateway at base URL https://api.kilo.ai/api/gateway, branded as "The Universal Gateway for AI Inference" and positioned as the routing layer that powers Kilo's own coding tools. It provides one key across 500+ models from 60+ providers, with pay-as-you-go billing at "exact provider rates" and no inference markup, plus BYOK where you "pay only for the routing, not the tokens". Its pricing page lists a Free access tier at $0/forever with access to default free models, basic analytics and BYOK support, and credit purchases carry a 5% processing fee with no subscription required.

No free tier stated on official pages (checked 2026-09-16): Kilo Gateway publishes a Free tier at "$0/forever" giving access to default free models, but no dollar or token credit allowance.

LLM.Kiwi is an AI model router offering access to various LLMs via an OpenAI-compatible endpoint. It provides a free tier with a rate limit of 40 requests per hour for specific high-resource models and higher limits for pro subscribers.

No public free-tier information found in the source reference; treat free-tier availability as unverified rather than confirmed absent.

LaoZhang AI is an enterprise-grade AI API integration platform reselling over 200 models from OpenAI, Anthropic, and Google. It provides an OpenAI-compatible interface, invoice support for businesses, and a small signup credit ($0.05) for testing.

No public free-tier information found (researched 2026-08-29).

LiteRouter is an AI API aggregator that provides access to hundreds of models from various providers via a single OpenAI-compatible interface. It supports free models with specific daily limits and pay-as-you-go billing for premium variants.

No public free-tier information found in the source reference; treat free-tier availability as unverified rather than confirmed absent.

Maritalk (by Maritaca AI) is a Brazilian AI provider specializing in Portuguese language models (Sabiá family). It offers an OpenAI-compatible API with inference hosted in Brazil for data sovereignty, featuring a daily free tier for chat and competitive PAYG token rates for the Sabiá-4 flagship model.

No public free-tier information found (researched 2026-08-29).

MegaNova AI is a comprehensive AI API aggregator and infrastructure provider, offering access to 200+ models including GPT-5.6, Claude 4, and proprietary Manta models. It features a tiered free quota system based on account status and spend, plus OpenAI-compatible serverless and dedicated GPU endpoints.

No public free-tier information found in the source reference; treat free-tier availability as unverified rather than confirmed absent.

MiniMax China (minimaxi.com) is the domestic-facing platform for MiniMax's foundation models. It mirrors the global platform's model offering and OpenAI-compatible API but optimizes for mainland China compliance and billing workflows.

No public free-tier information found (researched 2026-08-29).

Mixedbread AI is a specialized provider of knowledge-retrieval models and agents. Its platform offers an OpenAI-compatible API for its 'Toast' agent models, alongside managed search and indexing services billed by content tokens. It features a $5 one-time free credit for new users.

No public free-tier information found in the source reference; treat free-tier availability as unverified rather than confirmed absent.