These providers expose an OpenAI-compatible endpoint (openai_compatible: true), so migrating is usually a one-line change: keep the OpenAI SDK, swap base_url and api_key. Grab each provider’s exact base URL from its linked docs.
36 of 68 tracked providers match.
| Provider | OpenAI base URL | What's free | The catch | Verified |
|---|---|---|---|---|
| Google Gemini API (AI Studio) no cardno phonecommercial OKOpenAI-compat | https://generativelanguage.googleapis.com/v1beta/openai/ | Gemini 2.5 Flash, 2.5 Flash-Lite, 2.5 Pro (limited), embeddings, TTS models | Free-tier prompts/outputs may be used by Google to improve its products outside the UK/CH/EEA/EU. Since the 2026-03-23 terms, only Paid Services may serve API clients to end users in the EEA/CH/UK | 2026-10-08 |
| Groq no cardphonecommercial OKOpenAI-compat | https://api.groq.com/openai/v1 | Open-weight models (GPT-OSS, Qwen) plus Whisper, no credit card required | Limits apply at the organization level, not per API key. Phone verification required at signup | 2026-10-08 |
| OpenRouter no cardno phonecommercial OKOpenAI-compat | https://openrouter.ai/api/v1 | A rotating set of models with a :free suffix (16 on 2026-10-08; count fluctuates), single API across many providers | ToS (Jul 2026) prohibits reselling API access or building a competing service — platform-wide, not just the free models; per-model terms still apply | 2026-10-08 |
| Cloudflare Workers AI no cardno phonecommercial OKOpenAI-compat | https://api.cloudflare.com/client/v4/accounts/{account_id}/ai/v1 | 10,000 Neurons/day, all account plans | Resets daily at 00:00 UTC; overage on a Workers Paid plan bills at $0.011/1,000 Neurons. A few models (e.g. Kimi K2.6/K2.7-code, GLM-5.2) now require a Workers Paid plan | 2026-10-08 |
| Cohere no cardno phoneeval onlyOpenAI-compat | https://api.cohere.ai/compatibility/v1 | Trial (evaluation) API keys covering chat, embed and rerank | Explicitly for evaluation only — Cohere's terms prohibit production/commercial use on a trial key | 2026-10-08 |
| Cerebras cardcommercial OKOpenAI-compat | https://api.cerebras.ai/v1 | $5 in free credits for new accounts, usable across all public models | A verified payment method is required to activate Playground/API access (no charge until you buy credits). Credits expire 30 days after grant; whether any free access persists past expiry is not stated | 2026-10-08 |
| Mistral (La Plateforme) no cardphoneeval onlyOpenAI-compat | https://api.mistral.ai/v1 | "Free mode" (default for new accounts): API keys with included monthly usage and the lowest limits, "intended for evaluation and prototyping" | No credit card required. In Free mode Mistral may use your inputs and outputs to train its models unless you opt out in the Admin panel. Phone verification at signup was recorded in an earlier check; the current docs do not mention it. | 2026-10-08 |
| HuggingFace no cardOpenAI-compat | https://router.huggingface.co/v1 | Free CPU Basic + ZeroGPU for Spaces; Inference Providers has no included credit on the Free plan (credits must be purchased) — $2.00/mo is included only on PRO/Team/Enterprise | Credits only apply with "Routed by Hugging Face" billing, not with a Custom Provider Key | 2026-10-08 |
| SiliconFlow no cardphoneOpenAI-compat | https://api.siliconflow.cn/v1 | China platform (siliconflow.cn): several models permanently free at ¥0 (e.g. Qwen/Qwen2.5-7B-Instruct, tencent/Hunyuan-MT-7B, BAAI embeddings and rerankers). The international platform (siliconflow.com) lists no $0 LLMs and gives $1 in free credits instead | Free models live on the China platform and need real-name authentication to use all of them; accepted documents are a mainland ID card, HK/Macao or Taiwan travel permits, HK/Macao/Taiwan residence permits or a Chinese permanent-residence card for foreigners. Signup is by phone number; some virtual-carrier number ranges are blocked. | 2026-10-08 |
| Z.ai (Zhipu AI / GLM) no cardno phonecommercial OKOpenAI-compat | https://api.z.ai/api/paas/v4/ | GLM-4.5-Flash, GLM-4.7-Flash (text), and GLM-4.6V-Flash (vision) are officially listed as $0 cost (input, cached input, and output) on a permanent basis | Terms of Use prohibit using Z.ai to "develop, train, or enhance any algorithms, models, or technologies that directly or indirectly compete with us" — otherwise general use, including commercial, isn't restricted | 2026-10-08 |
| IBM watsonx.ai (Lite plan) cardOpenAI-compat | https://us-south.ml.cloud.ibm.com/ml/v1 | Lite plan: 300,000 tokens/month for foundation model inference, 20 CUH/month for ML tooling, 100 pages/month of document text extraction | Lite plan doesn't support fine-tuning of foundation or custom models; 1-day idle deployment timeout. Never expires or bills while inside quota, but a payment method (with a nominal ~$1 authorization hold) is required at signup | 2026-10-08 |
| OVHcloud AI Endpoints no cardOpenAI-compat | https://oai.endpoints.kepler.ai.cloud.ovh.net/v1 | Seven models are currently listed as Free: two Qwen3Guard moderation models, Stable Diffusion XL and four NVIDIA Riva TTS voices; access is available anonymously or with an API key tied to a Public Cloud project | European provider (France), relevant for EU data-sovereignty/GDPR-conscious use. The authenticated tier needs a valid payment method on the project (though "Free" models themselves don't charge); anonymous access needs neither an account nor a card. A separate general $200 Public Cloud trial voucher also exists but is unrelated to this free-model tier | 2026-10-08 |
| Fireworks AI no cardOpenAI-compat | https://api.fireworks.ai/inference/v1 | $1 trial credit | Fireworks uses prepaid credits. After the $1 credit is exhausted, add a payment method and credits (or enable auto top-up) to continue; account limits can rise with spend | 2026-10-08 |
| Baseten OpenAI-compat | https://inference.baseten.co/v1 | New accounts receive free credits; Baseten's current pricing page does not state the amount | Basic is $0/month, pay as you go. Current pricing confirms new-account credits but does not publish an amount or expiration date. Model APIs are priced per token; dedicated deployments are priced by compute time. | 2026-10-08 |
| Nebius AI Studio cardno phonecommercial OKOpenAI-compat | https://api.tokenfactory.nebius.com/v1 | $1 trial credit, valid for 30 days | Product renamed to "Nebius Token Factory"; a bank card is required to set up billing | 2026-10-08 |
| Novita AI no cardeval onlyOpenAI-compat | https://api.novita.ai/openai | $100 promotional credits valid for 90 days for eligible Novita Agent Sandbox usage (CPU, RAM, storage), granted after an account-setup survey; Novita does not say they apply to Model API calls; the pricing page also lists a few LLMs as Free (e.g. Ling 3.1 Flash, Apodex 1.1 Mini) | Signup advertises $100 Sandbox credits valid for 90 days; Novita's sandbox pricing doc describes them as credits for "eligible Novita Sandbox usage" and does not say they apply to Model API calls. No credit card required. Novita Terms of Service state that the Site and Marketplace Offerings may not be exploited for any commercial purpose without express prior written permission. | 2026-10-08 |
| AI21 Labs no cardOpenAI-compat | https://api.ai21.com/studio/v1 | $10 trial credit, valid 3 months | Card not required for the trial credit itself, required once it expires | 2026-10-08 |
| Alibaba Cloud (Model Studio) no cardOpenAI-compat | https://dashscope-intl.aliyuncs.com/compatible-mode/v1 | Typically 1,000,000 tokens per model for new users, Singapore region (International deployment scope) only | Covers real-time inference only; excludes batch invocation, fine-tuning, model deployment, custom models, PAI-DSW and OSS fees. Account information must be completed before activation. One quota per account (shared with RAM users); re-registering does not grant a new one. | 2026-10-08 |
| SambaNova Cloud commercial OKOpenAI-compat | https://api.sambanova.ai/v1 | Rate-limited free tier (applies when no payment method is linked) across the models listed in the Free Tier table | Free Tier applies when no payment method is linked to the account; SambaCloud ToS grants a commercial license (no evaluation-only clause). The previously-listed "$5 / 3 months" trial could not be re-confirmed on official pages (2026-07-30). card_required is unconfirmed because SambaNova's own pages disagree (read 2026-10-08): the rate-limits doc (https://docs.sambanova.ai/docs/en/models/rate-limits) says "Free Tier: Applied when there is no payment method linked with your account", while the plans page (https://cloud.sambanova.ai/plans) says "Add a payment method and purchase credits to run your first requests". Settling it needs a real signup, which is not done without the Owner's OK. The rate-limits doc (read 2026-10-09) also lists two preview models on the Free Tier at the same 20 RPM / 20 RPD / 200,000 TPD (DeepSeek-V3.2, gemma-4-31B-it), described there as having limited capacity and removable at short notice; they are not in models_free. | 2026-10-09 |
| Scaleway Generative APIs no cardOpenAI-compat | https://api.scaleway.ai/v1 | 1,000,000 tokens free + 60 min Whisper transcription; billing starts at token 1,000,001 | European provider (France). Free allowance is a one-time token bucket, not time-limited. Official rate limits apply once a valid payment method is registered; identity verification raises them. | 2026-10-08 |
| NVIDIA NIM no cardphoneeval onlyOpenAI-compat | https://integrate.api.nvidia.com/v1 | Trial credit, phone verification required | "Evaluation only, not production" per NVIDIA's own Trial Terms of Service — NVIDIA may discontinue the trial at any time with no continuity obligation | 2026-10-08 |
| Vercel AI Gateway no cardOpenAI-compat | https://ai-gateway.vercel.sh/v1 | Free tier: a monthly free credit covering a subset of models (Free Tier eligible models) at lower rate limits | A monthly free credit is stated in Vercel's docs (https://vercel.com/docs/ai-gateway/pricing, read 2026-10-09); the amount is not stated there. Once you purchase credits, your account moves to the paid tier and the monthly free credit no longer applies. BYOK is not available on the free tier. It routes requests to many providers rather than hosting models itself. models_free is a sample, not the full list: the first language model, by id, of each of the first eight providers in alphabetical order among the models flagged availableToFreeTier in the page data of https://vercel.com/ai-gateway/models, joined with type language in https://ai-gateway.vercel.sh/v1/models and excluding priority-tier variants (isFast) and the synthetic free copy (isSyntheticFreeTier) (131 such models from 30 providers, read 2026-10-09T03:00Z; the catalogue changes). We did not verify that the rendered ?freeTier=true filter matches that field. Five models priced at zero that spend no credit are tagged free in the API: convaiinnovations/laya, convaiinnovations/laya-free (evaluation type, not chat), inclusionai/ling-3.1-flash, inclusionai/ling-3.1-flash-free and poolside/laguna-s-2.1-free. | 2026-10-08 |
| Jina AI no cardno phonecommercial OKOpenAI-compat | https://api.jina.ai/v1 | 10M free tokens (one-time) across all models — embeddings, rerankers, classifier; plus a keyless Reader (r.jina.ai) for basic use | The 10M-token balance is a one-time grant that does not replenish; the keyless Reader is genuinely ongoing. Hosted API is commercial-OK and data is not used for training. No card required. | 2026-10-08 |
| AssemblyAI no cardno phonecommercial OKOpenAI-compat | https://llm-gateway.assemblyai.com/v1 | $50 free credit on signup (no card) — pre-recorded & streaming speech-to-text, Speech Understanding, and an OpenAI-compatible LLM Gateway (25+ models) | One-time credit, not renewing. Catch: the default model differs between free and paid accounts — set speech_models explicitly to avoid cost jumps on upgrade. Only the LLM Gateway is OpenAI-compatible. | 2026-10-08 |
| Clarifai no cardphoneOpenAI-compat | https://api.clarifai.com/v2/ext/openai/v1 | One-time $5 credit across serverless models (GPT-OSS-120B, Claude, Llama), vision, embeddings and image generation | Could not be re-verified on 2026-10-08: docs.clarifai.com no longer resolves and www.clarifai.com does not respond. Nebius has announced that Clarifai's core team joined Nebius and that it licensed Clarifai's inference technology. Treat this entry as likely discontinued. Previously recorded: SMS phone verification to claim a one-time $5 credit that expired 30 days after grant. | unverified |
| Arli AI OpenAI-compat | https://api.arliai.com/v1 | Free plan ($0): text LLMs for testing, max 5 requests every 2 days per model, 12K context, 1 request at a time, delayed responses; no image generation | Free tier is for testing only — very restrictive. Provider advertises zero-log / no data retention. Card requirement for the free tier is not stated on the pricing page. | 2026-10-08 |
| Ollama Cloud no phonecommercial OKOpenAI-compat | https://ollama.com/v1 | $0 Free plan: starter amount of cloud usage credits, limited to a smaller set of starter models; buying credits unlocks all cloud models | First-party — Ollama hosts the cloud models. Requires an ollama.com account + API key. The old 5-hour session / weekly limits are gone: usage is now credit-based with a monthly reset; Pro ($20/mo) includes $60 of credits. Prompts/responses are not logged or trained on. Whether a card is required is not stated on the current pricing page. | 2026-10-08 |
| ModelScope (API-Inference) no cardeval onlyOpenAI-compat | https://api-inference.modelscope.cn/v1 | Free API-Inference calls paid with daily 'Magicubes': 200/day for signing in + 50/day after linking an Alibaba Cloud account (do not roll over); calls cost ~0.5 / 1 / 2 Magicubes for lightweight / standard / flagship models | Alibaba's model hub — a different product from Alibaba Model Studio. Requires a ModelScope account bound to an Alibaba Cloud account with real-name (ID) verification — a practical barrier for non-China users. Explicitly non-commercial ("for developers to experience"). No card. | 2026-10-08 |
| Pollinations.ai no cardno phoneOpenAI-compat | https://gen.pollinations.ai/v1 | Quest Pollen: create an account, complete eligible Quests on the Quests dashboard and claim the rewards; Quest Pollen is spent before paid Pollen on regular (non-paid-only) models. No anonymous/no-key generation — every generation endpoint requires an API key and costs Pollen ($1 ≈ 1 Pollen). Text, image, video, audio, embeddings and 3D models behind one OpenAI-compatible API | The old anonymous 1 req/15s and Seed 1 req/5s tiers (and the watermark/nologo scheme) are gone from the current docs; the API moved to gen.pollinations.ai with keys from enter.pollinations.ai. Quest availability and reward amounts change ("the dashboard and the linked issue are the source of truth"). No credit card needed to try. Paid-only models require Paid Pollen. Commercial use is not explicitly addressed in the docs. The previous docs_url pointed at the stale master branch; the default branch is main. | 2026-10-08 |
| Moondream Cloud no cardOpenAI-compat | https://api.moondream.ai/v1 | $5/month usage credits in every workspace (Free plan) for the Moondream vision model — caption, query (VQA), detect, point | Recurring $5/month credit, no credit card required. Per the Terms (§5.2), data is not used to train Moondream's generally available models. Commercial terms for the hosted API not specified. OpenAI-compatible endpoint. | 2026-10-08 |
| Sarvam AI commercial OKOpenAI-compat | https://api.sarvam.ai/v1 | ₹100 in free credits on signup, usable across any API including chat completion (Sarvam-105B) and speech (STT/TTS); credits never expire | One-time signup credit shared across all APIs; credits never expire. Sarvam-M has been deprecated (use Sarvam-105B). Signup credits carry commercial production rights for generated Output. Card/phone requirement not stated in the docs. The docs list limits per plan but do not say explicitly that free-credit accounts are on Starter. India-focused provider (pricing in INR), strong Indic-language models. | 2026-10-08 |
| Tencent Hunyuan no cardOpenAI-compat | https://api.hunyuan.cloud.tencent.com/v1 | One-time free package on first activation: 1,000,000 tokens shared across Hunyuan text/vision models (Hunyuan-a13b, Hunyuan-role-latest, Hunyuan-translation, Hunyuan-translation-lite, Tencent HY Vision 1.5 Instruct, Hunyuan-turbos-vision, Hunyuan-t1-vision, Hunyuan-turbos-vision-video), plus a separate 1,000,000 tokens for Hunyuan-embedding | Free resource package valid 1 year from activation; unused tokens expire, and it does not roll into pay-as-you-go automatically. Activation requires Tencent Cloud real-name verification (enterprise or personal) — a practical barrier for many non-China users. Tencent says Hunyuan is migrating to TokenHub: the legacy platform will add no new models and stops supporting new model-service purchases; the legacy ChatCompletions and GetEmbedding APIs are marked pre-offline with expected shutdown 2026-12-21 (the OpenAI-compatible endpoint is the documented alternative), and the same API page (updated 2026-09-23) says the old Hunyuan console goes offline on 30 September. Whether new activations still receive the free package after the migration is not stated. Commercial-use terms not stated on the free-quota page. | 2026-10-08 |
| Poolside no cardOpenAI-compat | https://inference.poolside.ai/v1 | Free developer API key via Poolside Platform for Poolside's Laguna coding models (current lineup: Laguna S 2.1, Laguna XS 2.1, Laguna M.1); which models and limits apply to the free key is not documented | Coding-focused models. Docs describe Poolside Platform as "fast, free developer access"; paid access with a larger context window is via OpenRouter. The earlier 'free for a limited time' wording no longer appears in the current docs, and no end date is published. Self-serve key at platform.poolside.ai. No card was required at signup when last confirmed (2026-08); the current pages do not mention it. | 2026-10-08 |
| Upstage no cardcommercial OKOpenAI-compat | https://api.upstage.ai/v1 | $10 free credit on sign-up (per Upstage's Getting Started docs) for the API — Solar LLM (chat + embeddings) plus Document Parse / OCR / information extraction; Studio document agents include 10 free runs each | The Solar Pro 4 free window (2026-08-05 to 2026-08-11) has ended; standard pricing is $0.30 input / $1.20 output per 1M tokens from 2026-10-11 (UTC), after a 70%-off phase that runs until 2026-10-10. Beyond the signup credit, API use is pay-per-use, with prepaid commitment tiers from $100/mo adding bonus credit and higher limits. Studio agent runs need no card. A separate institutional grant (up to 1 year of free Solar Pro + Document Parse) exists for eligible schools, hospitals and nonprofits only. Solar Pro 2/3 and Solar Mini are deprecated on 2026-10-30 (KST). | 2026-10-08 |
| W&B Inference no cardcommercial OKOpenAI-compat | https://api.inference.wandb.ai/v1 | Serverless Inference credits on the Free plan "for a limited time" (amount not published); Free accounts have a default spending cap of $100/month | W&B Inference is now documented as CoreWeave Forge Serverless Inference (docs.wandb.ai redirects to docs.coreweave.com). The $100/month figure is a default spending cap, not a stated free-credit amount. When credits run out, Free accounts must activate pay-as-you-go on the Billing tab or upgrade; CoreWeave requires prepayment for paid access. Available only from supported geographic locations. OpenAI-compatible endpoint at api.inference.wandb.ai/v1; API keys are created at forge.coreweave.com. The pricing page lists "$5/mo free credit for a limited time" only under Pro and "Billed monthly" under Free, so the Free amount is unclear. | 2026-10-08 |
| Typhoon (SCB 10X) no cardno phonecommercial OKOpenAI-compat | https://api.opentyphoon.ai/v1 | Free to use research showcase API — all Typhoon models at $0 | By SCB 10X, the venture arm of Siam Commercial Bank, focused on Thai-language models. The free catalog spans LLMs (typhoon-v2.5-30b-a3b-instruct), OCR (typhoon-ocr family) and realtime Thai ASR (typhoon-asr-realtime, typhoon-isan-asr-realtime) — the hosted API is OpenAI-compatible, including audio transcriptions. Beta, provided as-is with no formal support; usage data is collected to improve the model; SCB claims no rights in outputs. Sign up for a free API key at opentyphoon.ai. | 2026-10-08 |
Quickstart — reuse the OpenAI SDK
Point base_url at the provider (URLs in the table above) and use its free key. Example: Groq.
from openai import OpenAI
client = OpenAI(base_url="https://api.groq.com/openai/v1", api_key="<YOUR_FREE_API_KEY>")
resp = client.chat.completions.create(
model="llama-3.1-8b-instant",
messages=[{"role": "user", "content": "Hello!"}],
)
print(resp.choices[0].message.content)curl https://api.groq.com/openai/v1/chat/completions \
-H "Authorization: Bearer $YOUR_FREE_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model": "llama-3.1-8b-instant", "messages": [{"role": "user", "content": "Hello!"}]}'