Free AI coding credits & free API tiers
What you can use for $0 to write code with AI, with the exact limits, base URLs and official links. Coding-relevant providers only.
Last verified 2026-10-19
Free API tiers
What's free: Free input and output tokens on many models, including gemini-3.8-flash, gemini-2.5-pro and gemini-2.5-flash. Not available for Gemini 4 Argon or Gemini 3.1 Pro Preview.
Limits: Per-project rate limits (RPM, TPM, RPD), shown in AI Studio; daily limits reset at midnight Pacific. Free-tier content is used to improve Google's products.
How to use: Create an API key in AI Studio, then point an OpenAI-compatible client at the base URL.
base_url = https://generativelanguage.googleapis.com/v1beta/openai/
Official link · source · verified 2026-10-19
What's free: Model variants whose ID ends in :free cost $0.
Limits: 20 requests/minute; 50 requests/day if you've bought fewer than 10 credits, 1,000/day after buying at least 10.
How to use: Use an OpenRouter key and a model ID ending in :free.
base_url = https://openrouter.ai/api/v1
Official link · source · verified 2026-10-19
What's free: Free API access to hosted open models such as openai/gpt-oss-120b and qwen/qwen3.8-27b.
Limits: Example (openai/gpt-oss-120b): 30 requests/min, 1K requests/day, 8K tokens/min, 200K tokens/day. Limits apply per organization.
How to use: Create a key in the Groq console and set the OpenAI client base URL.
base_url = https://api.groq.com/openai/v1
Official link · source · verified 2026-10-28
What's free: Every new Inception Platform account gets 100 million free tokens for the Inception API, with no payment details required. That includes Mercury 3, Inception's new diffusion model with a 1M-token context.
Limits: 100M tokens assigned once to each new account; Inception doesn't state an expiry. Free-tier rate limits per minute: 1,000 requests, 1M input tokens and 100K output tokens. After the free tokens, Mercury 3 costs $0.10 per 1M input and $0.35 per 1M output tokens at its 50% launch discount (regularly $0.20 and $0.70), with no end date stated for the discount.
How to use: Create an Inception Platform account, create an API key, and call the model mercury-3. Inception says the API is OpenAI API compatible, so coding tools that accept a custom OpenAI-compatible base URL can use it.
base_url = https://api.inceptionlabs.ai/v1
Official link · source · verified 2026-10-20
What's free: A small amount of free API credits for new users.
Limits: Amount not stated on the pricing page.
Official link · source · verified 2026-10-19
Free plans in coding tools and agents
What's free: Free model requests through Gemini CLI.
Limits: Gemini API key (unpaid): 250 requests/user/day, Flash model only. Google login (Gemini Code Assist for individuals): 1,000 requests/user/day per the quota table, but the same page says unpaid-tier users moved to Antigravity CLI on June 18, 2026.
How to use: Install Gemini CLI and sign in with a Gemini API key.
Official link · source · verified 2026-10-19
What's free: Copilot in the editor, CLI and agent mode.
Limits: 2,000 completions a month plus a limited allowance of GitHub AI Credits for chat and agents (GitHub doesn't publish the amount). Auto model selection only.
Official link · source · verified 2026-10-22 · pricing page
What's free: Codex in the ChatGPT desktop app.
Limits: GPT-6 Luna at Standard speed, subject to rollout.
Official link · source · verified 2026-10-19 · pricing page
What's free: Agentic IDE with a perpetual free tier.
Limits: 50 credits; open weight models and Claude Sonnet 4.5, rate limited.
Official link · source · verified 2026-10-19 · pricing page
What's free: Terminal and Warp Agent CLI.
Limits: Agent usage is pay-as-you-go, or bring your own inference.
Official link · source · verified 2026-10-19 · pricing page
What's free: All Amp features at $0.
Limits: Bring your own keys or ChatGPT/Claude subscriptions; no Amp token fees. Orbs pay-as-you-go, or run your own runners free.
Official link · source · verified 2026-10-19 · pricing page
What's free: Free, open-source coding agent for VS Code and CLI. Cline accounts also get rotating, limited-time free model promotions.
Limits: The extension is free; you pay only for inference, at cost or with your own API keys. Free promo models work up to a usage quota, only in the IDE extension and CLI (not the Cline API), and Cline says free usage may be used to improve the models. Cline said on X on Oct 22 that its new free models include Reflection Beam, Nemotron 3.5 Super VL, Poolside Laguna S 2.1, Nemotron 3.5 Lightning and Solar Mini 4. After the free quota: ClinePass ($9.99/month) or pay-as-you-go.
How to use: Sign in with a Cline account, set API Provider to Cline (or ClinePass) and pick a model tagged FREE.
Official link · source · verified 2026-10-23 · pricing page
What's free: OpenCode's own model console lists many models at $0 for a limited time, including DeepSeek V4.1 Flash Free, GLM-5.3-Flash Free, Beam Free, Flint 40B Free, MiMo-V2.6-Flash Free, Nemotron 3 Ultra Free, Nemotron 3.5 Lightning Free, Nemotron 3.5 Super VL Free and Big Pickle.
Limits: Free for a limited time, with no end dates in the docs. The docs say the DeepSeek and GLM free models follow zero data retention. Beam Free isn't used for training but is kept for 30 days, and isn't available in some countries, including China and Russia. Big Pickle, Flint, MiMo and the Nemotron free models may use your data to improve the model during the free period, and Nemotron 3 Ultra Free is trial use only (don't send personal or confidential data). LongCat 2.5 Preview Free and MiniMax M3.1 Flash Free have ended.
How to use: Sign in to the OpenCode console and copy your API key, run /connect in OpenCode and choose the OpenCode pay-as-you-go provider, then pick a model ending in Free with /models.
Official link · source · verified 2026-10-28
What's free: Qoder gives individual users a daily free allowance of DeepSeek-Flash (DeepSeek-V4.1-Flash-1017) at 0 credits, including at max reasoning, in the Qoder IDE, JetBrains plugin, CLI, Cloud Agents, mobile and web.
Limits: Limited-time event: Oct 20 to Oct 31, 2026, 24:00 UTC+8 (Oct 31, 4 PM UTC). Ends tomorrow, Oct 31. For individual Qoder International plans (Free, Pro Trial, Lite, Pro, Pro+, Ultra), not Teams or Enterprise. No claim needed and it works with a zero credit balance. Paid plans get more free uses; Qoder doesn't publish the daily amount. It's DeepSeek V4.1 Flash, not DeepSeek V4 Pro.
How to use: Update Qoder, sign in with an individual account and pick DeepSeek-Flash in the model selector.
Official link · source · verified 2026-10-30
What's free: Kilo Code, an open-source coding agent for VS Code, JetBrains and the CLI, offers free models through its Kilo Gateway, including an Auto Free router (kilo-auto/free) and models labeled (free) such as Recursive's Flint and NVIDIA Nemotron models.
Limits: No fixed quota published; some free models may be rate limited by the upstream provider, and the free model list changes over time. Kilo warns that Auto Free may route to providers that log prompts and outputs and use them to improve their services (including NVIDIA's free endpoints), so don't send personal or confidential data. Autocomplete and background tasks use credits by default unless you switch them to a free model.
How to use: Sign up for Kilo Code, then pick Auto Free (kilo-auto/free) or type free in the model picker and choose a model labeled (free). In the CLI, run kilo and use /models.
Official link · source · verified 2026-10-23
What's free: Freebuff is a coding agent (desktop, web and terminal) funded by text ads. Its Free plan costs $0 with no card and includes models such as DeepSeek V4.1 Flash, GLM 5.3 Flash and MiMo v2.6 Flash.
Limits: Free plan: 30 hours of usage refilled daily, unlimited messages within each one-hour session. Models use hours at different rates (for example about 10 hours a day of DeepSeek V4.1 Flash, or 2 hours of Claude Haiku 5.5). Hours reset at midnight local time and don't roll over. These allowances are for full-access countries; elsewhere, or on a VPN, you get limited mode with a smaller allowance and fewer models. Freebuff shows ads, says model costs and allowances may change during its beta, and shows each model's data-use notice before you start, so read it before using private code.
How to use: Download Freebuff (desktop, web or CLI), sign in with GitHub or Google, and pick a model in the model picker. Check the model's data-use notice before using private code.
Official link · source · verified 2026-10-22
What's free: Mistral's agent on web and mobile.
Limits: Limited messages and web searches; limited coding sessions.
Official link · source · verified 2026-10-19
App builders with Free plans
- Lovable: 5 build credits/day (up to 30/month) + 20 Cloud credits/month + 4 AI-feature credits · source
- Bolt: 300K tokens/day, 1M tokens/month; Bolt branding; up to 333k web requests · source
- v0: $5 one-time credits; 7 messages/day; deploy to Vercel; GitHub sync · source
- Base44: 25 message credits/mo (up to 5 a day), 100 integration credits/mo, up to 5 apps · source
Not listed: Qwen coder models on Alibaba Cloud (official pricing page didn't load) and the Mistral API free tier (official doc unavailable).
Third-party free model hosts
Free models served by resellers, routers and hosts, not by the model makers. Read the warning first. Full details, privacy quotes and setup steps: third-party free model hosts.
⚠ Read this before you use them
- These are resellers and hosts, not the companies that make the models.
- Your prompts and code pass through their servers.
- Don't send private or client code, passwords, API keys or other secrets.
- Free capacity can queue, change or end without notice.
- Read each host's terms and privacy policy before you sign up. We quote the key lines below.
- Quotadex isn't affiliated with any of these hosts and earns nothing from these links.
| Host | Free models | Limits | Card needed? | Last checked |
|---|---|---|---|---|
| IteraCompute Reseller / open-model host | qwen/qwen3.8-27b (Qwen3.8-27B) | 30M tokens/month (input + output, incl. reasoning), 160,000 tokens/min, 128K context, output capped at 8,192 tokens. Resets on the 1st at 00:00 UTC, no rollover. Free accounts may queue at busy times; no uptime SLA. | No card needed | 2026-10-30 |
| BeatAPI Reseller / API aggregator | glm-5.3-flash-free (GLM 5.3 Flash), deepseek-v4.1-flash-free (DeepSeek V4.1 Flash), gemini-3.8-flash-free (Gemini 3.8 Flash) + 1 more | 1 successful request per minute per account before your first top-up, up to 10 per minute after any top-up. No published context or output limit for the free IDs, and no promised daily allowance. glm-5.3-flash-free and minimax-m3.1-flash-free don't support /v1/responses; minimax-m3.1-flash-free can take tens of seconds to return its first token. | No card or top-up needed | 2026-10-30 |
| OpenRouter (:free models) Router (forwards to many providers) | recursive/flint-v1.0:free, xingchen-agi/xing4.0-29b-a4b:free, nvidia/nemotron-3.5-super-vl:free + 13 more | 20 requests/minute. 50 requests/day if you have bought fewer than 10 credits, 1,000/day after buying at least 10. A negative balance can block free models too. | No purchase needed for the 50/day tier (card not stated) | 2026-10-30 |
| Groq (Free tier) Inference host for open models | openai/gpt-oss-120b, openai/gpt-oss-20b, qwen/qwen3.8-27b | Groq's table lists, for openai/gpt-oss-120b and qwen/qwen3.8-27b: 30 requests/min, 1K requests/day, 8K tokens/min, 200K tokens/day. The page also says your exact limits are on the Limits page of your account, and its wording about whether this table is the Free or Developer baseline is unclear. | No card for the Free tier (a payment method is asked only to upgrade to Developer) | 2026-10-30 |
| Ollama Cloud Cloud host for open models | beam (Reflection Beam, free for a limited time) | Ollama's $0 pay-as-you-go plan has no subscription and allows 3 concurrent requests; other cloud models are billed per token. Beam has a 1M-token context and isn't in Ollama's per-token price list. Ollama's Beam post says it's free for a limited time but gives no end date or token cap, and doesn't say whether Beam needs bought credits. | The pay-as-you-go plan is $0 with no subscription. Ollama doesn't say whether a card is needed. | 2026-10-30 |
| NVIDIA API catalog (build.nvidia.com) Model host / trial API (NVIDIA hosts its own and other companies' models) | deepseek-ai/deepseek-v4.1-flash (DeepSeek V4.1 Flash), z-ai/glm-5-3-flash (GLM-5.3 Flash), nvidia/nemotron-3-ultra-550b-a55b (Nemotron 3 Ultra) + 1 more | Up to 40 requests/min and 10,000 requests/day, per NVIDIA's model pages, which add that limits may vary by model and that traffic from other users may cause throttling. Trial use only: NVIDIA's trial terms say not to use it in production. | NVIDIA says Developer Program members get free access to NIM API endpoints for prototyping. It doesn't say whether a card is needed. | 2026-10-30 |
Links are not affiliate links. Last checked 2026-10-30.
FAQ
Which AI APIs have a free tier for coding?
Google's Gemini API (via AI Studio) has a free tier on several models, OpenRouter offers free :free model variants, and Groq has a free plan for hosted open models. Anthropic gives new API users a small amount of free credits.
Can I use these free APIs in coding tools?
Yes, where the tool supports a custom OpenAI-compatible endpoint: set the base URL shown on each card and use your own key. Cline, for example, supports your own API keys.
Are free tiers private?
Not always. Google says content on the Gemini API free tier is used to improve its products. Check each provider's terms.
How often is this checked?
Each card shows the date we checked the official page. We re-check weekly.
API vs subscription calculator · Student discounts & deals
Never get surprised by a price hike
The free weekly Quotadex changelog: price changes, new limits, and which plan is the best deal right now.
1 email a week. No spam. Unsubscribe anytime.