Try dolphin-mistral-24b-venice-edition free on LLM EU
This card is served by OpenRouter, with Not verified: the provider has not documented this claim. and 128k tokens, recorded under Not verified: the provider has not documented this claim.. A new organization starts with $5 of credit, so a first request against dolphin-mistral-24b-venice-edition is paid for by that grant rather than by a card.
Try dolphin-mistral-24b-venice-edition with $5 of free credit on a new organization: OpenAI-compatible API, recorded price, and zero retention of prompt and completion bodies.
$5.00
- Price
- $0.20 / $0.90 per million tokens
- Region
- global
402 budget_exceeded · llmeu.reason = insufficient_credit
- Provider
- OpenRouter
- Parameters
- Not verified: the provider has not documented this claim.
- Context window
- 128k tokens
- License
- Not verified: the provider has not documented this claim.
- Sovereignty class
- non_eu_third_country
- Retention
- Runs outside the EU. Reachable only under residency "any".
Terms
Promotional credit. Every new organization receives USD 5.00 of promotional inference credit. It is a marketing benefit, not a purchase: it is not cash, it has no cash value, it cannot be withdrawn, transferred, sold or refunded, and it is not redeemable for money or for any other service. It expires 90 days after the account is created and unused credit is forfeited without compensation. It can be spent only on inference through this API and is consumed before any other balance; its amount, availability and eligibility may change or be withdrawn at any time. The models it reaches are the endpoints your own policy allows, and those endpoints may be operated by third parties; the price recorded per endpoint governs how far the credit goes, so USD 5 buys a different number of tokens on different models. When you name a specific model this service serves that model or refuses the request by name — it never answers with a different model presented as the one you asked for — and every response states which model and which region actually served it. Promotional credit carries no guaranteed availability, no service level and no priority over any other traffic.
With llmeu-auto the router chooses the model under your policy. Name a model and the service serves that model or refuses it by name — never a different one presented as the one you asked for. Every response states which model and which region actually served it.
llmeu.model_used · llmeu.inference_region · llmeu.route_reason
How to make the first request
Create an organization, then mint an API key in the console. The credit is granted once with the account, and the key spends it after the same policy checks every other request passes. Send one POST to /v1/chat/completions with the model id dolphin-mistral-24b-venice-edition and a bearer key, or point an OpenAI SDK at the base URL. The playground sends the same request if you would rather not write code first. The response carries the standard OpenAI fields plus a block naming the trace id, the region, the sovereignty class, the retention that applied, a cost estimate and the sentence that explains the route.
curl https://api.llmeu.com/v1/chat/completions \
-H "Authorization: Bearer $LLMEU_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"dolphin-mistral-24b-venice-edition","messages":[{"role":"user","content":"Hello"}]}'What dolphin-mistral-24b-venice-edition costs
The price is not set on this page: it is read per million tokens from the endpoint row that serves dolphin-mistral-24b-venice-edition, with $0.20 recorded for input and $0.90 for output. A rate the row does not record is shown as unverified rather than filled in. Requests are routed to a row that is live and allowed by your policy, and the response returns the actual usage with a cost estimate in USD; the trace keeps the same figures in USD and EUR. The $5 grant is spent at whichever of those rates applies.
The limits around the free credit
The default policy caps spending at 10.00 USD per organization per day; reaching it answers 402 budget_exceeded before any endpoint is scored. No card is on file, the credit does not auto-recharge and it is granted once per organization. Your policy also bounds which models, providers, regions and data classes a request may use, and a model this deployment does not serve is refused by name. Those are limits, not promises: the credit pays only for what the policy already allows.
Frequently asked questions
Is dolphin-mistral-24b-venice-edition really free to try?
The first five dollars are. A new organization is granted 5.00 USD of credit once, and dolphin-mistral-24b-venice-edition can be called with it under the same conditions as any other call: a live endpoint has to serve it and your policy has to allow it. After that balance is gone the API refuses the request rather than billing you, and no card is on file that could be charged.
What licence applies to dolphin-mistral-24b-venice-edition?
The licence the card records, with the link to the text its publisher publishes — nothing on this page reproduces or replaces it. The same licence governs whether the weights may be used commercially, which is why a card whose licence forbids commercial serving stays catalogue-only here. The free credit does not change a single condition of that licence: what you may do with dolphin-mistral-24b-venice-edition is decided by that document, not by the balance you hold.
What happens when the $5 credit is used up?
The balance reaches zero and the API refuses the next request instead of serving it, naming the limit that stopped the call. Nothing is silently charged and nothing is served anyway. Your account, keys, traces and usage history stay, and an owner or admin can add credit in the console — there is no automatic recharge. If dolphin-mistral-24b-venice-edition is not reachable after that, it will not be reachable with credit either: balance never overrides a policy rule or an endpoint that is absent.