Schedule a Call
As of October 5, 2026, OpenRouter has 17 free models, every model whose ID ends in :free, plus the openrouter/free auto-router. That’s down from the 27+ this list covered in July, so the table below shows every model that’s actually free today, grouped by what it’s good at. Search “best free OpenRouter models” and you’ll still find lists naming model IDs pulled months ago, like DeepSeek or Gemini “free” endpoints that no longer exist. Copy one of those old IDs into your code and you’ll just get an error. Every model below was checked against OpenRouter’s own live model list, so here’s what’s actually free right now, and which one to pick.
What “Free” Actually Means on OpenRouter
“Free” means $0 per token, no hidden per-request fee, and no credit card needed to start. But free doesn’t mean unlimited.
OpenRouter caps free (:free) models at 20 requests per minute, always. Your daily limit depends on your account: 50 requests a day if you’ve bought less than $10 in credits in total, or 1,000 a day once you’ve bought $10 or more — a limit that sticks even if you later spend that balance down. One catch: a negative balance can trigger a 402 error even on :free models. Think of it like a gym guest pass: free entry, but the front desk still caps your visits.
Older posts often quote “200 requests per day.” That’s outdated; OpenRouter’s rate limits documentation confirms the 50/1,000 split above.
All 17 Free OpenRouter Models (October 2026)
This is the full free roster from OpenRouter’s public model list, pulled on October 5, 2026, grouped by what each model is built for. Context windows are for the free endpoint, which can be smaller than the paid version of the same model. Two free endpoints already have an end date: both Poolside Laguna models are scheduled to expire on October 31, 2026, and Dots3-Note Preview on December 31, 2026.
That’s all 17. Popular July picks that are no longer free include Owl Alpha, Laguna M.1, gpt-oss-120b, gpt-oss-20b, GLM-4.5-Air and Kimi K2.6; some are still on OpenRouter as paid models, others are gone. The roster rotates often, so check OpenRouter’s live models page before building anything that depends on one specific ID. For a coding-only shortlist, see our best free OpenRouter models for coding.
Which Free Model Should You Actually Use?
Skip the full list and go straight to a pick:
- Building a coding agent? cohere/north-mini-code:free for something that will stick around, or poolside/laguna-s-2.1:free while it lasts (free until October 31, 2026).
- Need something fast and light on tokens? nvidia/nemotron-3.5-lightning:free or google/gemma-4-26b-a4b-it:free, both with only a few billion parameters active per token.
- Working with huge documents or long agent chains? nvidia/nemotron-3-ultra-550b-a55b:free, nvidia/nemotron-3.5-lightning:free or thinkingmachines/inkling:free, all with a 1M-token context window.
- Need image or vision input? qwen/qwen3.8-27b:free or google/gemma-4-31b-it:free; for audio too, thinkingmachines/inkling:free or nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free.
- Don’t want to choose at all? Set your model to openrouter/free — OpenRouter’s own auto-router. It picks a free model that matches what your request needs, kind of like ordering “chef’s choice”: you’re not picking the exact dish, but the kitchen picks something that fits what you asked for.
Myths Still Floating Around
Two things worth correcting first:
“DeepSeek, Gemini, and Mistral have free models on OpenRouter.” Not currently true: none of the three has a :free model as of this October 2026 check, despite guides still pointing to old free IDs. Google’s free options on OpenRouter are its open Gemma 4 models, not Gemini.
“Free models cap out at 200 requests a day.” Outdated. The real split is 50/day or 1,000/day, depending on whether you’ve bought $10 in credits.
How Many Free Models Are There, Really?
Honestly, it depends on the day you check. On October 5, 2026, OpenRouter’s own model list has exactly 17 free models, down from 27+ in July — the roster shifts constantly as providers add, pull, or reprice models. Treat 17 as a snapshot, not a fixed number.
If you’re building something long-term, don’t hardcode a model ID and forget it. Check the free filter on openrouter.ai/models before you ship, and keep a fallback ready in case your pick gets pulled.
Quick FAQ
- Do I need a credit card to use free models?
- No. Sign up with just an email, and you can start using :free model IDs right away.
- Can I use free OpenRouter models in production?
- You can, but it’s risky — no uptime guarantee, and rate limits bite under real traffic. Best for prototyping, not anything customer-facing.
- What happens if a free model gets removed?
- Your code keeps running, but requests fail or start billing you if that model switches to paid. Build in a fallback, or route through openrouter/free.
- What’s the fastest way to try one without writing code?
- Open any model’s page on OpenRouter and use its built-in chat playground — no API key setup needed just to test.
- How many free models does OpenRouter have in 2026?
- 17 as of October 5, 2026: every model ID ending in :free in OpenRouter’s public model list, plus the openrouter/free router. The count changes often, so re-check the Price: Free filter before you ship.
- Is DeepSeek free on OpenRouter?
- Not right now. Every DeepSeek model on OpenRouter is paid as of October 2026, though some cost only a few cents per million input tokens.
What to Do Next
Your one action right now: pick one model from the table, open its page on OpenRouter, and try it in the playground. Two minutes, no cost, nothing to break.
For the full walkthrough — getting your API key, sending your first request, avoiding rate-limit errors — BULDRR AI’s guide to free OpenRouter API keys covers it step by step, and best free OpenRouter models helps you pick one for everyday tasks. And if you’re testing these models inside an automation setup instead of raw code, most drop straight into an n8n workflow through the OpenRouter node, so you can swap models without touching your underlying logic. Our guide to using OpenRouter with n8n shows how.
Each button opens the assistant with the prompt already written, asking it to teach this article step by step, in simple words.
ChatGPT
Claude
Grok
Perplexity
Add as a preferred source
