> OpenRouter
OpenRouter is a unified, OpenAI-compatible API aggregating hundreds of model providers behind one key. Free model variants (:free IDs) are genuinely free with documented platform caps — 20 requests/min and 50/day for accounts that have purchased under 10 credits, 1,000/day once you've purchased 10+ — while paid variants have no platform request cap. Fallback routing, per-key credit caps, and BYOK are built in.
// pros
- One API key reaches hundreds of providers — switching models is a model-ID change, not a new integration.
- Free variants (:free) cost nothing, with documented platform caps: 20 req/min, 50 req/day under 10 credits purchased, 1,000 req/day at 10+.
- Automatic fallback routing retries other providers for the same model when one is rate-limited or down.
- Per-key credit caps plus a GET /api/v1/key quota endpoint make spend budgetable per integration.
- BYOK: route through OpenRouter with your own upstream keys, tracked as separate usage.
// cons
- Free variants are hard-capped at 20 requests/min, and a negative account balance blocks even free models with a 402.
- Pricing is pass-through from upstreams: the same model costs different amounts by provider, and rates move when upstreams move them.
- The auto-router has no fixed rate of its own — it bills at the underlying model's price, so costs are unpredictable until you pin a provider.
- A 429 can originate at OpenRouter or at the upstream provider; you must read error metadata to tell which.
The practical free tier to build on: prototype on :free variants, and buy 10 credits early to lift the daily cap from 50 to 1,000 requests. For anything you depend on, pin the provider and set per-key credit caps.
written by Mike Fleming, Uprouter editorial · verified against official sources September 16, 2026
How we estimated this: the free allowance is converted to a dollar value using the provider’s own pay-as-you-go rates for the models a typical developer would reach for. When the allowance is metered in requests, tokens or minutes rather than dollars, we assume median usage of the cheapest capable model. Because assumptions are involved, the figure carries the est. flag — treat it as an order of magnitude, not a quote.
note: :free models cost $0, no card; 20 req/min; 50 :free req/day with <10 credits, 1000/day after buying $10+ credits (verified FAQ; failed reqs count). ~$1.5/mo equivalent at 50 req/day, ~1K tokens/call.
| Plan | Type | Monthly | Input /1M | Output /1M | Note |
|---|---|---|---|---|---|
| Free tier (source-reported) | free | — | — | — | :free models are free and require no credit card; rate-limited to ~20 requests/minute, with 50 :free requests/day if you hold <10 credits, rising to 1000/day once you buy >=10 credits (failed requests also count). |
| Free tier | free | — | Free | Free | 50 req/day; 1,000 req/day with $10 lifetime top-up |
| Pay-as-you-go | payg | — | $1.25 | $10.00 | Example for Anthropic Claude 3.7. 5.5% platform fee applied to listed rates. |
Pricing normalized from public sources — always verify with the provider.
// model prices
| Model | Input /1M | Output /1M | Context |
|---|---|---|---|
| Free Models Router | Free | Free | 200k |
| nvidia/nemotron-nano-9b-v2:free | — | — | — |
| z-ai/glm-4.5-air (:free) | — | — | 131k |
| qwen/qwen3-next-80b-a3b-instruct:free | — | — | 262k |
| Gemma 4 31B | — | — | 262k |
| gpt-oss-120b | — | — | 131k |
| gpt-oss:20b | — | — | 131k |
| Llama 3.3 70B Instruct | — | — | 131k |
| Nemotron 3 Ultra | — | — | 256k |
| Qwen3 Coder | — | — | 262k |
| Auto Router | — | — | 2M |
$ Does OpenRouter have a free tier?
Yes — OpenRouter offers a free tier that we value at roughly $1.50 in free credits (estimate — derived from rate-limited usage at pay-as-you-go prices). Always confirm current limits on OpenRouter's official pricing page — free tiers change without notice.
$ How much does OpenRouter cost per 1M tokens?
The cheapest model we track at OpenRouter is Free per 1M input tokens and Free per 1M output tokens. This is normalized from OpenRouter's published pricing — verify with the provider before purchasing, since prices change frequently.
$ Is OpenRouter safe to use?
Uprouter rates OpenRouter as low risk. Low risk as an aggregator: OpenRouter is a transparent intermediary — published rate-limit policy, per-key spend caps, and it never holds your upstream keys unless you opt into BYOK. The real exposure is upstream dependency: availability and price are only as good as the provider serving that model at that moment, which is precisely what the built-in fallback routing exists to manage. Risk ratings are editorial, evidence-based, and never influenced by affiliate relationships.
Raw community sentiment, shown unblended. Votes never feed into objective data fields like price, risk, or status — and never into sort order.
No published reviews yet — be the first.
// more_informationprice history · live status · risk & confidence · code snippets
price_history
No recorded changes yet.
live_status
full history →best-effort probes · not a guarantee · last 19h ago
| Checked | Status | Latency |
|---|---|---|
| 19h ago | up | 80ms |
| 2d ago | up | 78ms |
| 3d ago | up | 94ms |
| 5d ago | up | 72ms |
| 6d ago | up | 80ms |
| 8d ago | up | 92ms |
| 9d ago | up | 73ms |
| 11d ago | up | 127ms |
| 12d ago | up | 75ms |
| 13d ago | up | 77ms |
| 14d ago | up | 65ms |
| 15d ago | up | 80ms |
| 16d ago | up | 61ms |
| 16d ago | up | 60ms |
| 17d ago | up | 79ms |
| 18d ago | up | 60ms |
| 19d ago | up | 64ms |
| 19d ago | up | 66ms |
| 20d ago | up | 66ms |
| 21d ago | up | 95ms |
| 22d ago | up | 69ms |
| 22d ago | up | 242ms |
| 23d ago | up | 79ms |
| 24d ago | up | 80ms |
| 25d ago | up | 65ms |
| 26d ago | up | 72ms |
| 27d ago | up | 77ms |
| 27d ago | up | 73ms |
| 28d ago | up | 82ms |
| 29d ago | up | 77ms |
Uprouter runs its own probes; best-effort, not a guarantee. Probe results depend on our network location and cadence — for production, always use the provider’s own status page.
risk_&_confidence
Low risk as an aggregator: OpenRouter is a transparent intermediary — published rate-limit policy, per-key spend caps, and it never holds your upstream keys unless you opt into BYOK. The real exposure is upstream dependency: availability and price are only as good as the provider serving that model at that moment, which is precisely what the built-in fallback routing exists to manage.
Confidence reflects how much verifiable evidence (official docs, probe history, community reports) backs this entry. Risk is our editorial assessment of operator reliability and terms — not financial advice.
code_snippets
# OpenRouter's own API — get a key from their dashboard
curl https://YOUR-PROVIDER-ENDPOINT/chat/completions \
-H "Authorization: Bearer THEIR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "nvidia-nemotron-nano-9b-v2-free",
"messages": [{ "role": "user", "content": "Hello" }]
}'Direct to the provider — Uprouter receives nothing; verify their current pricing before purchasing.
Ranked by neutral similarity signals only — live status, free-tier shape and free-credit proximity. Commercial relationships never influence this list.