Intelligence with Everyone: RL @ MiniMax, with Olive Song, from AIE NYC & Inference by Turing Post

Hi. Hi, everyone. I'm Olive.

It's my great honor here today to present on our new model, MiniMax M two. I actually lived in New York City for six years, so it feels great to come back, but with a different role. I currently study reinforcement learning and model evaluation at MiniMax.

Let me just get a quick sense of the room. Who here has heard or have tried of MINIMAX before? Oh, couple of their yeah.

Not everybody, but I guess yeah. But here's the value, right, of me standing here today. So we are a global company that works on both foundation models and applications.

We develop multi mod modality models including text, vision language models, our video generation model Halo, and speech generation, music generation stuff. We also have, many applications including agents and stuff, in house. So that's the specific thing that's different from the other labs for other companies.

So we both develop foundation models, and applications. So we have research and developers sitting side by side working on things. So our difference would be that we have first hand experience from our in house developers into developing models that developers would really need in the community.

And here I want to introduce our MiniMax M2, which is an open weight model, very small with only 10,000,000,000 active parameters, that was designed specifically for coding a workplace agentic task. It's very cost efficient. Let me just go over the benchmark performance because people care about it.

So we rank very top in both intelligence benchmarks and also agentic benchmarks. We, I think we are on the top of the open source models. But then numbers don't tell everything because sometimes you get those super high number models, you plug into them, into your environment and they suck.

Right? So we really care about the dynamics in the community. And in our first week, we had the most downloads.

And also we climbed up to top three token usage on open router. We're very glad that people in the community are really loving our model, into their development cycle. So today, what I wanna share is how we actually shape these main model characteristics that made M2 so good in your coding experience.

And I'm gonna present to you, the training behind it that supports each one of them. From coding experience to long horizon state tracking tasks, to robust generalization to different scaffolds, to multi agent scalability. So first let's talk about code experience, which we, which we supported with, scaled environments and scaled experts.

So, developers need a model that can actually work in the language they use and across the workflow that they deal with every day. So which means that we need to utilize the real data from the internet and then, scale the number of environments so that the model when during training, for example, reinforcement learning, it can actually, react to the environment, it can actually target verifiable coding goals and to learn from it. So that's why we scaled both the number, of environments and also our, infrastructure so that we can perform those training very efficiently.

So, with data construction and reinforcement learning, we were able to train the model so that it's very strong, it's full stack multilingual. And what I wanna mention here is that besides scaling environment that everybody talks about, we actually scale something called expert developers, as reward models. So as I mentioned before, we have a ton of, super expert developers in house that could give us feedback to our model's performance.

So they participated closely into the model development and training cycle, including problem definition. For example, bugs, bug fixing, for example, repo refactoring and stuff like that. And also they identify the model behaviors that developers enjoy and they identify what's reliable and, what developers would trust.

And they give precise reward and evaluation to the model's behaviors to the final, deliverables so that, it is a model that developers really want to work with and that can add efficiency to the developers. So with that, we were able to lead in many, languages in real use. And the second characteristic that Minimax M2 has is it performs good in those long horizon tasks.

Those long tasks that require interacting with complex environments that requiring, using multiple tools with reasoning. And we supported that with the interleaved thinking pattern, and reinforcement learning. So what is interleaved thinking?

So with a normal reasoning model that can use tools, it it normally works like this. You have the tools information given to it, you have the system prompts, you have user prompts, and then the model would think and then it calls tools. It can be a couple of tools at the same time, and then they get the tool response from the environment, and then it performs the final thinking and deliver final content.

But but here's the truth. Right? In real world, the environments are often noisy and dynamic.

You can't really perform this one test just by once. You can get, tool errors for example, you can get, unexpected results from the environment and stuff like that. So, what we did is that we imagined how humans interact with the world.

We we would look at something, we get feedbacks and then we think about it. We think if the feedback is good or not and then we make other actions, make other decisions. And that's why we did the same thing with our M2 model.

So if we look at this chart over a diagram on the right. So instead of just stopping, after one round of tool calling, it actually thinks again and reacts to the environments to see if the information is enough for it to get what it wants. So basically, we call it interlead thinking or people call it interlead thinking because it interlead thinking with tool calling.

A couple of times it can be, you know, tens to a 100, turns of tool calling within just one user interaction term. So, it helps, adaptation to environment noise. For example, just like what I mentioned, the environment is not stable all the time and then something is suboptimal and then it can choose to use other tools or do other decisions.

It can focus on long horizon has, can automate your workflow, using for example, Gmails, Notions, Terminal all at the same time. You just need to maybe make one model call with minimal, human intervention, it can do it all by itself. And here's a cool illustration on the right because it's in New York City, I feel the vibe of trading and marketing.

So you can see that there was some, there was some perturbations in the stock market, I think last week. And then our model was able to keep it stable. So just like I said, there's like environment noise, there's new information, there's like, yeah, news.

It looks like there's like other trading policies and stuff like that, but our model was able to perform pretty stably in these kind of environments. And the third characteristic is our robust, generalization to many agent scaffolds, which was supported by our perturbations in the data pipeline. So we want our agent to generalize.

But what is agent generalization? At first, we thought it was just tool scaling. We train the model with enough tools, various tools, kind of new tools, we invent tools, and then it would just perform good on unseen tools.

Well, that was kind of the truth. It worked at first, but then we soon realized that if we perturb the environment a little bit, for example, we change another agent scaffold, then it doesn't generalize. So what is agent generalization?

Well, we conclude that, it's adaptation to perturbations across the model's entire, operational space. If we, think back, what's the model's, operational space that we talked about? It can be tool information, it can be system prompts, it can be user prompts, they can all all be different.

They can be the chat template, they can be the environment, they can be the tool response. So what we did is that we designed and maintained perturbation pipelines of our data so that, our model can actually generalize to a lot of agent scaffolds. And the fourth characteristic that I want to mention is the multi agent scalability, which is very possible with M2 because it's very small and cost effective.

I have a couple of videos here. This is M2 powered by our own MiniMax agent, app. We actually have a QR code down side, so if you want, you can just scan and try it.

It's like an Asian app we developed. And here, we can see different copies of M2. It can do research.

It can write the research results and analyze it and put it in a report. It can put it in some kind of front end illustration and they can work in parallel. So because it is so small, and so cost effective, it can really, support those long run authentic tasks and tasks that maybe, require some kind of parallelism.

So what's next? Right? For Minimax M two, from what I've introduced, we gathered environments, algorithms, data, expert values, model architecture, inference, evaluation, all these stuff to build a model, that was fast, that was, intelligent, that could use tools, that generalizes what's next.

For, M 2.1 and M three point in the future, we think of better coding, maybe memory or context management, proactive, AI for workplace, vertical experts, and because we have those great audio generation and video generation models, maybe we can integrate them. But all our mission is that we are committed to bring all these resources, whatever is on the screen and maybe more, yeah, and values, and put them all together to develop models for, the community to use.

So, we really need feedback from the community if possible because we want to build this together and this is kind of a race that everyone needs to participate. And then we are committed to share it with the community. And that is all the insights for today.

We really hope, again, we really hope you to try the model because it's pretty good and then we can contact contact us up there. You can try the models by scanning the URL code. Yeah.

Basically, that's it. Thank you all for listening. Georeanforcement learning model tries best to hack a lot of things.

The current open models can achieve that level of understanding. It is a solvable problem and we are working on it. Engineering is very, very, very important.

I didn't know that during school.

at 9PM on Sunday night. Does everyone work like this at the company? I'm really impressed.

I think different people work on different schedules. We do have people who work even overnight, but they sleep at daytime. So we have a very flexible schedule.

It goes with your experiment. For example, if the experiments run for all day, the person can take a break. And then if there are a lot of analysis to do, maybe because we are very curious about the results and we're very passionate, right?

We can't really wait a very long time. So yeah, everyone has their own schedule. That's telling about the success of the models.

I think that's influenced that. You specialize in reinforcement learning and model evaluation, as far as I understand, which are two of the least forgiving parts of model development. And you also have more constraints than big AI labs.

What does a good day look like for you and what does a bad one look like? I can share something about our recent weeks. So there's not a whole good day or a whole bad day.

We were joking that during one day we have good results in the morning and then sometimes it becomes bad results at night. Sometimes we call it, we have like ICU in the morning and then KTV at night. So typically a good time would be usually receiving some good results or like, even if running into new problems is a good time.

So if we, for example, during reinforcement learning, right, we can see the model doing a lot of different stuff to achieve the results. And sometimes we just discover like new model behaviors. And that's really exciting, even though it might not be safe or like it might not be expected, it's kind of exciting.

So like, I call it a good time. A bad time would be, there really isn't a bad time except for, you know, looking at the, you know, just finding out the bad results. The moment itself is bad, but then trying to figure out the problem and breaking it down is pretty good time.

What were the recent model behavior that you didn't expect? So the model, the reinforcement learning model tries its best to hack a lot of things, right?

How do you do that? Frankly speaking, I'm not the expert person on that part. We have a whole team on role playing her stuff.

I'm not an expert, but we do have a lot of discussions. We do believe that role playing or like, you know, accompanying human or like human interactions is very important in the life of AI or like how it would change our social life in the future. And it absolutely represents some ability that's very superior because that's like human like, you know, it has emotions where you understand your emotions.

It's not just, you know, working out some exams. That's absolutely another side of the AI capability.

intelligence with everyone. What does it mean for you? For me personally, I feel like it's more like how it changes my life and it enables me to do more work and then how it can connect me better to different people.

Because for example, before I wouldn't be able to understand a lot of very professional, for example, right? Very professional coding problems or like optimization problems. And now I am able to do that with AI and so that I can communicate with more people and exchange more ideas.

That's one side. And on the other side, it generally helps my daily life, right? So it helps with my work, my daily routine, my self care, changes life for me.

And I hope that it changes life for everybody, But obviously in a good way. Can you tell me a little bit how day to day work is organized in your lab? I remember from your talk at AI Engineer that it's very interconnected between developers and researchers.

I would love to hear more about that. Absolutely. We sit around every day.

So we share our experiment results. For example, as I just said, during experiments, for example, reinforcement learning experiments, we see some scores going up high. We look at the model's behaviors.

We look at the model behaviors with the developers in that area as well. We sit together, and then they will spot the issue right away. And then we we are able to come up with new ideas to fix it or, like, building more data on it.

the current version, what are the biggest problems you're trying to solve comparing to the previous version? One important thing we focus on right now and also in the future is human alignment. Because we are focusing on coding models for 2.

1, 2.2, and the m two series. Right?

And then what we realized is that for it to become very productive in our daily work or, like, for it be productive and safe at the same time, we have to do a lot of alignment on it. So the model can't just grow on its own and then do some dangerous behaviors just to achieve the final goal. So for us, the important thing would be how we define human alignment, how we define expert expectation, how we actually train the model to be more aligned with our expectations.

So, want to go in some real details here.

algorithm. So we have the theoretical reinforcement learning algorithm. And then, but when we implement it, it could be a little bit off that creates a little bit gap to the theoretical extreme of this algorithm.

So that's how we think and approach to this problem is that we try to, you know, scale to the theoretical extreme. And for example, the precision part is one thing that we found that would, kind of like prevent us from being close to that extreme and that's how we solve it. That was a very funny story actually for when we discovered that.

I kind of talked about it when we published, Minimax M1. So during our experiments, we found that the accuracy didn't go up. We looked at like layer by layer.

We looked at the log props layer by layer and found it out. Theoretically speaking, it has to work, right? And then there has to be some gap between the theoretical and, how we approached it.

So we thought about the gap and we'll analyze it layer by layer and then eventually found it. Is there anything like this happening now? Definitely.

Yeah. Every single day, every single day and in every different groups. I can't actually disclose something that we haven't had the concrete conclusion because we want our conclusions for like anything public to be very concrete and we understand it very deeply.

So if we have, you know, breakthroughs, we'll definitely publish it later. But I'll have to say we do encounter these problems every day and we think I think it's called the first principle, right? So we think from the very fundamental part of the problem and then approach it.

The models that you launch are open weights.

expert in building developments or building things with models. I feel like because it's open way, people can have free reuse with it. For example, they can deploy it by themselves or if they can even fine tune it with the Waze and then have all the data on their properties.

Something that's sensitive safety or something that's like alignment safety. We have that as our evaluation. And then before launching about one or two weeks before launching, do scaled up evaluations and we do scaled up alignments on the model.

And that's how we assess if the model is safe. But then if it's already open weight in the wild, people actually can do something on it. I guess that's what you are approaching at, right?

People can do more things on the model that we can't control. I don't know how we handle that. Frankly speaking, there are laws on that, right?

There are regulations where people do agree on some moral standards on that.

collect feedbacks on the model itself first. So when we publish a model outside, many developers use it or, like, many people use it. We collect it systematically.

We analyze each problems. Some of them are fundamental. Some of them are, you know, just something that we missed and we can fix it real quick.

So there are two parts. First, we do the internal evaluation with the developers and they point out problems and that's how we can fix this part, but they are not enough. And more feedbacks will come to us after we officially publish the models and then we collect it.

Because, the way we organize our group is that different people work on different capabilities of a general model. If we collect some things that we think we should improve in the future, different people take their parts, right? So they're like, okay, I think I can solve this issue and I'll solve it next in the next generation.

like coding? I think we are approaching the generalized models. It's just that we are putting more emphasis on coding.

For example, our model also you can take it into any general agent scaffold, including our own agent product. And that's for general purpose. We do work on like, you know, researching, report writing, PPT, stuff like that.

That's more general. Personally speaking, I feel like with coding, you can, you know, kind of structure the whole world or like you can model a lot of stuff with coding. Yeah.

Engineer it. Yeah. With engineering.

So behind it, it's it's scaled up humanity for me. So it itself has a lot of intelligence with it and a lot of work to do. So that's how we view this issue.

But we do work on generalized stuff and even more generalized stuff in later versions. For example, our model can do some general workplace scenarios in the future and that's not just coding.

extraordinary infrastructure. For example, for reinforcement learning, right? The very important issue besides algorithm, besides things that people have been working on for a very long time, what's special for AgenTeq stuff is how we define agents, how we define how agent model would work.

First, you need to define the task, you need to define the model's goal, especially in a long horizon task, you need goals that are actually hard and diverse. And then the second part is that you need environments. You need great engineering environments, scaled up environments, different diverse environments, not just coding or like more, for example, workplace, different kinds of tools.

That's great engineering. And then you need great infrastructure. You need outstanding RL infrastructure to let the model really roll out in a very long horizon.

And then with, you know, very efficient, for example, GPU use very efficient training, rollout in training and stuff. I feel like that's what's different in agentic reinforcement learning as compared to before. Are you affected by GPU's constraint?

How do you solve the compute problem? We do have a team that works on how we utilize the computes the most. You know, that's actually one of the RL scaling issues, is to utilize the compute very efficiently.

So their purpose would be to minimize the compute use while training more. Personally speaking for me, I don't really have a GPU constraint.

How to use them more efficiently? Or it's just that team? We are actually the same team because we're actually the reinforcement learning team.

We view this issue on different perspectives. It can be implementation, right? It can be implementation.

You can view it as a data perspective. You can view it at different perspectives. But our goal is the same.

We are actually working on some new agentic reinforcement learning stuff, but it won't really come out with 2.2. With the next generation model, are actually we are still working on it.

I'm not sure what I can share or not. So I can share it later when I have concrete conclusions, as I said before. I can't really say something that we don't document yet.

Will it be available when the model is out? That depends on our time.

What's the cons to that? So for our team, like for the researchers team, we always wanted to go open source because open source communities is fantastic. I learned that from day one when I joined the team is open source community is fantastic.

So as researchers, we did want to join the open source. But then on the other hand, speaking of the cons, we are a company that people can care about if this can sell money or like, if this is a business. So the cons would be, if the ways are open source less people will use APIs, But then as a researcher, that really isn't my focus that much.

So I'm not very confident about the company strategy. For the tech part, we just believe that we can build better models with the open source community. How much do you use open source tools yourself from different other companies?

A lot. For example, inference, we use I'm not sure if I'm allowed to say specific open source branches, but then we collaborate with both VLM and SGLON. They are open source code repositories.

How do you look at it? For example, there are a lot of open source agent scaffolds, both coding agents and agent agent scaffolds that we use ourselves to test our models. And then we look at their logics.

We look at their code to see how the design specific both scaffolds and for example, engines. And then we take what they worked on was really good. And then we reflect on how we think the problem, how we structure the problem, if we're on the same page and stuff like that.

We learn from each other.

up and it's a different compute and you need to have a talent for that to use it, engineering talent, instead of just like, you know, choosing closed API, turn it on and use it. Do you have any difficulties to that or just inside the company, the open source stack is like established and working? Personally, I don't have a problem with that.

There are other open source models. And if they publish, I'll just download it and and, you know, deploy it on a machine and then work with it if I want. Personally, I don't have that issue.

But if there are, you know, personal developers out in the world, I understand the problem, especially when they don't have their own compute. Then it will be easier to connect to a model through, for example, open router and stuff like that. Do you use a lot of other open models on the the same open router, let's say?

Do you play with them? Yeah. Play with them.

I would play with them day one. If they release at midnight, I play with them in midnight.

I don't actually take notes, but I do have my personal evaluation stack. A list of their fun questions that I like to test with every single model to see how they work. Can you tell me about it?

That's super interesting. Yeah. I've been collecting a bunch of questions since I entered the company on different areas, including logical reasoning, including mathematics, proofs, including report writing, including agentic tasks and stuff like that a lot.

And some of them are very, I just like to see how the model reacts to these problems and how they approach it.

if it works well. From the professional evaluation perspective, five questions isn't enough. So if we want to do like, you know, very standard and very fair comparison among models, we have make it very, a confident test.

So there has to be a certain number of questions in each domain to see how the model performs. And usually you need to test it multiple times because models are not very stable themselves. If you are testing for fun, use the fun questions.

But if we are actually assessing the model's capabilities, we need some sets and that's very fair among different models. That's correct because some problems are not correct. Some questions, the answers are not single, for example.

And sometimes when we run the test, the environments are not fixed, for example. For example, the goal you had just wouldn't pass and stuff like that. So if we're doing professional evaluation, we have to make sure the evaluations is correct.

It's diverse. It's above a certain threshold so that the test is, you know, confident. You mentioned characters.

How do you work with your model character? I don't work on my model's characters. Okay.

That's how I think of this issue. A general model should have all characters or like it should be able to perform all characters. It might have default character.

If the user want it to be a different character, it should be. If the model injected into the system prompt, it should be. That's how I view this issue.

I find it hard to adjust the new models because they're so different in terms of character all the time. I just don't even understand why it happens. I think it has to be something related to the data that the model was trained on.

The different patterns the models have been trained on. And also different people different team might have their own, you know, constitution. In the system prompt?

Like, there are all those risks for open model. Where does it break first? I think open models are not very good at adjusting to different environments from what I see right now.

We can see, for example, cloud. Right? People use cloud in different coding environments, and then people think they perform well in all environments where like different tool definitions and stuff.

But I don't feel like the current open models can achieve that accuracy or like that level of understanding of the different environments. Why? Where is the problem?

I don't know how Cloud does it, but for me, I think it is a solvable problem and we are working on it. We are improving it in 2.2, but it's still not as good as, for example, Opus.

But for 2.5, it might be. We do have some systematic research going on in the area that has shown some results now, but it's still, it's not concrete conclusion.

So I won't say it. So curious. But do you think it's the problem with compute because they have this infinite amount they can just throw at it?

I feel like compute is one side, but how we structure the problem and how we approach it is another side. And that's where we are more confident on that we can solve the issue.

if it's launched by the time that the interview is out? Can you give me some overview? Better coding, obviously, and better multilingual coding, obviously, and more stable than before.

It has better performance in the area of 2.1. In different areas, it's better.

More stabilized, longer horizons, and stuff like that. And we are testing it in different environments right now, and we believe that it's better than before. So different coding environments, right?

interest that helps you kind of cross pollinate with what you do? Can you tell me how you stay up to date and what inspires you? There are different articles, different box going out every single day and a bunch, you know, all the information.

How we deal with it is that we have a internal agent that tracks all the new articles and blogs and papers, and then it dispatches to different subjects and then it summarizes then it analyzes and researchers. We have an internal researcher, if I call it, that does some filtering by itself and then it gives what is filtered to us and then we can improve the researchers if we think it doesn't do well. And that's how we filter out a lot of information first.

And then we play with new code repositories with coding agents so that we can understand it more quickly and then play with it more quickly. So we're pacing up with all improvements with agents and with our models for our models.

I joined the team, I thought I would be reading papers every day because that's what I was doing during school. Right? During a lab.

We would read papers, come up with ideas, implement ideas, run experiments. If the experiments results are good, we run it as a larger scale. I thought I was about to do that.

But then what I realized was that when joining the company and then working for a couple of months, you already become pretty much top of the area or like of the industry and that you have to come up with something that's really new or like you encounter problems that you just don't know how how to to solve. It's not like you can read a lot of papers and then building up your thinking on the papers. It's more like you need to really understand the problems from the fundamental and then think of it from the fundamental so that you can find the right solution.

And another things would be that engineering is very, very, very important. I didn't know that during school because during school or during labs, it's more toys as compared to companies. It's not that scaled up.

But when you do scale up data, you scale up compute, scalable people, right, you encounter engineering issues that you need to tackle very beautifully, and engineering is very important. That's part two that was different from what I imagined. Pretty much these two, I feel like.

We have a meta goal at the company level. For example, we want to improve the AI's capabilities improving productivity, for example, because that's how people view. So we have a company's mission.

As a single researcher in the team, we have our own missions that we set our own goals with. What is your goal currently? For the next generation, I would be able I really want the model to to be working elegantly with experts.

So it's more like better collaboration with experts, with developers. That's my goal as well. But that's maybe like two versions away, I think.

I think we're launching one version about per month or a month and a half. Right? For longer horizon, are we are definitely working on it.

But then for me, for the goal that I set along that path, that's like a three month away thing. But for the better collaboration thing, that's like a one month or two months away thing.

That has some overlaps with defined concept of continual learning. And by overlap, I mean, I think both conceptually and technically. But I don't feel like they are exactly the same or like the things that I talked about at the summit was not at the level of the full continued learning.

It's more like on the path to that. How do you see it being solved? Any ideas?

We do think that's a different problem definition or that's a different way of the model working with people. And we are working on that norm with our own defined question. But if I need to say how we approach it, I would say too, we would approach it through experiments.

That's a very interesting question on continued learning. It's very still very exploratory. Right?

That's definitely what we are going at. But then it has different phases or different stages.

outlined the stages? Outlining stages, we do have our internal definitions that I didn't prepare today. I would say first would be to be more stabilized into long horizon task and what I said at the summit.

Right? And then the next thing would be optimization.

environment feedbacks in a new environment. It needs to know what to explore and what environments to see, because it's a partially observed environment. It needs to know the actions that it takes to receive better information and then better reactions, and then perform harder, more complex tasks in the environments.

That's like more of stage one, right? That's pretty simple. Basically, all agented models can do that to some extent, maybe not perfectly, but to some extent.

And that's how we can actually solve it with our current algorithms. But we do see different norms of how the model improves itself in the environment that we don't have a concrete conclusion yet. Maybe in 2.

5 we will. That will be a different definition than what I said. The model itself would be defining its own goal.

That's something that would be different. Thank you so much. My last question is about AGI.

and ASI every day. Actually, when I was interviewing with MiniMax, when I was interviewing with our CEO, I said the same thing because he asked me the same thing. Right?

And then what I said was that I think people talk about AGI. People have different definitions of AGI, but we can only know the definition of AGI when we achieve it. Or like, it is still progressing so fast that the definition even changes every day and people have different comments on it.

But what I think is more important is we actually work towards it, work towards our own definitions of AGI. As long as we figure it out, it becomes true. And that's what I said during the interview, and that's still my view today.

The definition will become true when it becomes true. When we see it, we know it's CGI.

Something like The Art of Creativity or something that I read during undergrad. So it's a long time. I don't remember the exact name.

Yeah, there is a book. The name Art of Creativity. How did it influence you?

It opened up how I think of my own mind a lot and then how I view the world and how I view problem solving. For me now, problem solving is more of discovery. That's how I would summarize it in one quote.

Thank you so much. Thank you for your time. That was very interesting.

Thank you for having me.