Transcription
Hey everyone, welcome back to another new exciting video. Guys, I have found one free inference and its name is Open Inference, and it is currently giving access to the world's top popular models like Deepseek Version 3.1, Cohere 3/40 billion, and also this Kim 2 and GPT OSS from OpenAI. You will get access for free, and there is no rate limit.
Now, you will tell me that on the OpenRouter platform, also this Cohere 3 Coder and Deepseek Version 3.1 free models are available, right? Like this one is free from OpenRouter. Then what advantage does this Open Inference provide?
See, if you use this Cohere Coder through OpenRouter and if you are using this free model, see here in brackets they have written "free." Then there is a rate limit, meaning it is not completely free to use. They are just giving access for free for some kind of requests. Like if you go to their blog post here, you see that on OpenRouter, if you are using any free model, then you can do 20 requests per minute. But if your quota got completed, then you have to purchase one plan. And if you purchase the plan also, there are also this 50 model requests per day you can make. Okay.
But if you use this Open Inference, and then through this Open Inference, if you use this Cohere 3 Coder model, then there is no rate limit, and there is no money you have to pay, and you can use it unlimited. Okay. So, what is the advantage? I think you got the point.
Now let me show you that how you can enable this and how you can access this. Here you see this is the post introducing Open Inference. And guys, one thing, I am not promoting this product. I just found this helpful, so that's why I just thought to share it with you, and that's why I am making this video.
Here you see, "Introducing Open Inference: Free access to the best open-source AI models on OpenRouter. Built to accelerate open research. Usage is locked, anonymized, and shared to benefit the entire community. Since we are driven by data, not money, we prioritize inference quality over quantity, and it is very possible that our data and research will contribute to all future models. Want your voice to shape what comes next."
So, this is the website, openinference.xyz. I have just given this link in the description. Go there, you will find this kind of interface. Okay.
Now, to use this, here you see that "How to activate Open Inference." They have given this link, so just click on this link: openrouter.ai/settings/privacy. And yes, they are collaborating with OpenRouter to give this access. So, if you go to this link, there you will find this kind of page in your OpenRouter account. And here you see, this is the thing: "Enable free endpoints that may publish prompts." So, just turn it on. Okay.
So, if you just turn it on, what will happen? That whatever the prompts or whatever the data that you are sending to this model, they will use that as a dataset. Okay. So, that means they will use your data for training purposes, and that is the permission that you need to give to this Open Inference, and now they will give you the access to these models for free. Okay.
Now, what is the rate limit here? You see that, let me show you. One person actually asked this that thing in this post. Here you see, here you see, "What is the personal usage limit?" 500 requests per day. Okay. There is no money you have to give. You can do 500 requests per day, and I think that is enough. Okay.
Now, through OpenRouter free model, here you see that they were providing 20 requests per minute, and this is the per day limit. Here you see that 50 requests per day. Okay, 20 requests per minute and 50 requests per day. And this Open Inference is providing 500 requests per day, and this is enough, right? And here you see that is there any token kind of things? No. Here you see that "We allow 500 requests per day." So, the amount of token doesn't matter. What amount of token you are consuming, it doesn't matter. So, they are actually providing the access based on the number of requests, and currently, it is 500 requests per day. Clear? I think it is clear.
And here you see that, now you will ask me that, can I use it for business purpose or commercial purpose? No, you cannot use it for commercial purpose. You can use it for your personal use only. And here you see that, "Basically, we are limiting usage to personal use only. So, anything that looks like commercial or large-scale usage gets blocked. Also, personal use is very efficient to serve due to prefix caching."
And here, one thing you have to remember that if you are thinking that you can use this free API in your product and it will, and it will, you will sell your product to the people, then in that case, if they find that many people are using your product and many requests are going to their server, so in that case, they can block your account. So, remember these things. It is not for the commercial purpose. It is only for the personal use only.
Now, the thing is that you have to understand. So, what we have done? We have just turned on this feature, "Enable free endpoints that may publish prompts." When you turn it on, what will happen? Now, if you use this Cohere 3 Coder free model, it will be used through this Open Inference only, not the OpenRouter. Just think this way. Okay.
So, if you don't turn it on, okay, if you don't turn it on, where was that? Here you see, if you don't turn it on, it will use the by default OpenRouter. But if you turn it on, now if you use this model, Cohere 3 Coder, or Deepseek Version 3.0 free model, or Kimik K2 free model, then it will use the Open Inference. Okay. And now you will get the 500 requests per day. There is no rate limit you will get. Okay.
So, how to use that? Here you see, go to your client or root or whatever you are using. So, just go there and here in this section, just select this OpenRouter. Yes, you have to select this OpenRouter. Provide your API key. I think many of you already know how to get the API key. So, let me show you again also. Just go to this OpenRouter settings, API key, or if you click on your profile, here also you will get your API keys. And here, click on this "Create API key," and here give any name here, click on this "Create," and your API key will be created. Just copy that and paste it here in this section.
And in this model section, let's say I want to use the, I want to use the Cohere 3 Coder. Okay. Cohere 3 Coder because in my last video, already I have discussed that Cohere 3 Coder is based. Okay. So, where is the Cohere 3 Coder? Yes, this is the Cohere 3 Coder free model. So, if you use this, then it will not show you the rate limit. Okay. You can do the 500 requests per day.
Now, this will be very much helpful because those who are actually using these open models for learning purposes or for research purposes, for them, they will get the advantage to use these models for free, and also they can continue their work without any rate limit. Okay.
So, I hope that these new things will be helpful for you. And also, if you have any question, just let me know in the comment section. And if you are visiting this channel for the first time, then don't forget to subscribe to this channel. Don't forget to like this video. Also, if you found this helpful, this detailed explanation, transparent explanation will be helpful for you guys. And please watch the other videos also, like Gemini Live Instant Update, and this GitHub Copilot, Claude, Kimik's New Agent Mode, okay, Computer GPT, Codex Plus, Winder. So, please watch the other videos. Okay.
And another thing is that if I go to this Open Inference. Okay. So, if you go to this OpenRouter and if you search this Open Inference, here you will get this Open Inference page. And here you see that these are the models currently they are giving access. But I think that later they will add more models to their Open Inference. And here you see that many people are accessing these models. This is the graph. Okay. And, and yes, if you click on this Kimik K2, then you will see that it will take you to that free version of the model. Okay. So, Cohere 3 Coder, it will take you to that free version of the model, and this Deepseek V3, that free model, and GPT OSS, that is the free model. Okay.
So, see you guys in the next video. Thanks for watching. Bye-bye.