Drishan Arora (@drishanarora) / X

Drishan Arora

30 posts

Drishan Arora

@drishanarora

AI Researcher @DeepCogito

San Francisco, CA

Joined February 2013

Drishan Arora
@drishanarora
Apr 8, 2025
Today, we are launching @DeepCogito, where we are building general superintelligence. We are also releasing open models of 3B, 8B, 14B, 32B, and 70B sizes trained using our research on iterated distillation and amplification (IDA). From evals so far, each model outperforms the
433K
Drishan Arora
@drishanarora
Jul 31, 2025
Today, we are releasing 4 hybrid reasoning models of sizes 70B, 109B MoE, 405B, 671B MoE under open license. These are some of the strongest LLMs in the world, and serve as a proof of concept for a novel AI paradigm - iterative self-improvement (AI systems improving themselves).
452K
Drishan Arora
@drishanarora
Apr 8, 2025
Replying to @drishanarora
All models we create will be open sourced. More details in the blog post:
Deep Cogito
From deepcogito.com
19K
Drishan Arora
@drishanarora
Apr 8, 2025
Replying to @drishanarora
We use IDA to remove the intelligence ceiling. Simply put, use more computation to let the model arrive at a better solution, and then distill the expensive thinking process to the model's own parameters. As the LLM improves in intelligence, the thinking process itself becomes
40K
Drishan Arora
@drishanarora
Apr 8, 2025
Replying to @drishanarora
Traditional LLMs are upper bounded in intelligence by their overseers (larger teacher models or human capabilities). Building superintelligence requires not only matching human-level abilities but also to uncover entirely new capabilities we have yet to imagine.
19K
Drishan Arora
@drishanarora
Apr 8, 2025
Replying to @drishanarora
The Cogito v1 models can be downloaded on @huggingface or @ollama, and can be accessed via API through @FireworksAI_HQ or @togethercompute. Each model can answer directly (standard LLM), or self-reflect before answering (like reasoning models).
22K
Drishan Arora
@drishanarora
Apr 8, 2025
Replying to @drishanarora
From what we can tell, we're still in the early stages of this scaling curve - IDA is incredibly powerful and generalizes across domains. Most notably, our 70B model also outperforms Llama 4 Scout (109B MoE) distilled from a 2T model. As we improve and iterate on our
18K
Drishan Arora
@drishanarora
Jul 31, 2025
Replying to @drishanarora
The models are built on our work on building superintelligence using Iterated Distillation and Amplification (IDA). In particular, we scale the model’s intelligence prior by the model internalizing the reasoning process using iterative policy improvement, rather than simply
15K
Drishan Arora
@drishanarora
Jul 31, 2025
Replying to @drishanarora
More details in the blog post:
Deep Cogito
From deepcogito.com
12K
Drishan Arora
@drishanarora
Jul 31, 2025
Replying to @drishanarora
The Cogito v2 models can be downloaded on @huggingface, and can be accessed via API through @TogetherAi or @runpod or @baseten. You can also run them locally using @UnslothAI. Each model can answer directly (standard LLM), or self-reflect before answering (like reasoning
14K
Drishan Arora
@drishanarora
Jul 31, 2025
Replying to @drishanarora
This seems to be a novel scaling paradigm where the models develop more “intuition”, and serves as a strong proof of concept for self-improvement. Since the Cogito models develop a better intuition of the trajectory to take while searching at inference time, they have 60% shorter
14K
Drishan Arora
@drishanarora
Jul 31, 2025
Replying to @drishanarora
We plan to scale up and expect significant improvements by self improvement over the next few months using this approach. From experiments so far, this technique is far more efficient than simply “searching more” by longer reasoning chains. All models we create will be open
13K
Drishan Arora
@drishanarora
Aug 1, 2025
A small update - we had more traffic than anticipated. However, the endpoints are now scalable on Together AI for all models, including the 671B MoE. Test out the model here: together.ai/models/cogito-… (A huge thanks to the folks at @togethercompute for making this happen so
Drishan Arora
@drishanarora
Jul 31, 2025
Today, we are releasing 4 hybrid reasoning models of sizes 70B, 109B MoE, 405B, 671B MoE under open license. These are some of the strongest LLMs in the world, and serve as a proof of concept for a novel AI paradigm - iterative self-improvement (AI systems improving themselves).
Cogito v2 preview - 671B MoE API | Together AI
From together.ai
23K
Drishan Arora
@drishanarora
May 29, 2025
Data quality has the biggest impact on LLM performance - far more than most algorithmic improvements. As we build models that move towards superintelligence, the paradigm for LLM evaluation would need to evolve to increase the strength of the overseer. Very excited to see this -
Mannat Sandhu
@mannatsan
May 29, 2025
Today, we are launching Anthromind, where we are building scalable oversight for AI systems. As LLMs and AI systems grow more intelligent, the data needed to evaluate, supervise and align these models requires higher intelligence, often surpassing human expertise. Traditional
2.5K