OpenAI releasing new open model in coming months, seeks community feedback

For those wondering how to answer &quot;what do you want to see from an open model&quot; I put this in: an open weights end to end multimodal model, a model large enough to act as a functional teacher along with a range of nicely distilled smaller sizes, code repo to make training &#x2F; finetuning easy.<p>As I write this, I&#x27;d also like to request a set of tool calling LLMs in various sizes. Feels to me like a small fast local tool calling assessor with a large context to support a lot of MCP functions would be very useful locally.

Easy. A reasoning model with better performance than QWQ but at 21B (like Reka Flash 3) and good tooling call support. A model as “intelligent “ as Qwen2.5 but personality and creativity of Gemini (or Gemma at a minimum)

Open model is such a misnomer: it&#x27;s like calling an ELF an &quot;open executable&quot;.<p>Distributing things for free doesn&#x27;t make them &quot;open&quot;. The reality is that free (as in <i>free beer</i>) weights are closer to freeware than anything open source. In fact, since all these models are build using pirated media a more appropriate term could be plain old <i>warez</i>.

This definitely feels like a move made out of desperation, but I admit I&#x27;m curious to see what they release.<p>It&#x27;s fascinating that some of the best-funded startups of the 2020s are all rushing to commoditize their core technology as quickly as possible. They all seem to have the same business model as the Change Bank from SNL

Current open-weights models are not as multilingual as GPT3 or GPT4. I&#x27;d like to see support for more languages.

Q: How open?<p>Can it be used commercially? Is the training protocol going to be open? Or just the weights released like Llama models?

&gt; Let&#x27;s start with your details

OpenAI should release its frontier model as an open-weight model. There are already open-weight models that match OpenAI&#x27;s best models (at the time of their release), so the idea that OpenAI would lose something by making its frontier models open-weight doesn&#x27;t hold up. With an open-weight model, they would instantly kill any proprietary competitor, similar to what Google did to all its competitors with Android.<p>IMO, OpenAI should focus on tooling, infra, and setting standards for AI apps and profit from those. MCP is what &quot;custom GPT&quot; is supposed to be; OpenAI lost that battle, among many others.<p>Gopher started as a freely available protocol but later changed its licensing terms, requiring fees to be paid—similar to what OpenAI has done. We know how Gopher ended up: today, most people haven&#x27;t even heard of it, despite its adoption at the beginning.

What&#x27;s the major use case of open source models? The stable diffusion community seems pretty active. A lot of fine tuning to generate NSFW. What about LLM?

Don’t kid yourself about the intentions. It’s only so enterprise customers can deploy a shitty chatbot on premise for their &quot;secret data&quot;. The models will be free for commercial use until 1 mil turnover or something like that.

<i>What would you like to see in an open-weight model from OpenAI? Explain what you would use it for</i><p>See... that&#x27;s kinda the idea behind an open model. I don&#x27;t <i>have</i> to explain what I would use it for.