Poetiq, an emerging AI company focused on enhancing the reasoning capabilities of large language models (LLMs), has announced that it has raised $45.8 million in Seed funding. The round was co-led by FYRFLY Venture Partners and Surface Ventures, with participation from Y Combinator, 468 Capital, Operator Collective, Hico Ventures, and Neuron Venture Partners. This funding milestone comes shortly after Poetiq delivered standout results on ARC-AGI-2, a widely respected industry benchmark for machine reasoning and progress toward artificial general intelligence (AGI).
Notably, Poetiq’s performance on the ARC-AGI-2 Public Evaluation Set (December 23, 2025) demonstrated the strength of its AI meta-system. Unlike traditional approaches that rely on extensive fine-tuning or costly reinforcement learning, Poetiq works alongside any frontier LLM including OpenAI’s ChatGPT, Anthropic’s Claude, Google’s Gemini, and Meta’s Llama to help models learn faster and solve more complex problems. Instead of requiring thousands or even millions of examples, Poetiq enables clients to provide just a few hundred samples. As a result, its system generates specialized agents that continuously improve accuracy while reducing operational costs.
AI Authority Trend: Jelou Raises $10 Million to Build AI Apps That Move Money on WhatsApp
Founded in June 2025, Poetiq is led by co-CEOs Shumeet Baluja, PhD, and Ian Fischer, both former AI researchers at Google DeepMind. Baluja previously served as CTO of Jamdat Mobile and spent over two decades at Google DeepMind, where he helped establish foundational research initiatives and contributed to more than 170 patents across neural networks and machine learning. Fischer, meanwhile, joined DeepMind through its acquisition of Apportable, where he was co-founder and CTO.
During their time collaborating at Google DeepMind, the founders observed a fundamental challenge: frontier LLMs often struggle with both simple and complex reasoning tasks. Moreover, the prevailing solution large-scale pre-training and post-training using reinforcement learning can take weeks and remains prohibitively expensive for most organizations.
“LLMs are impressive databases that encode a vast amount of humanity’s collective knowledge,” said Shumeet Baluja, co-CEO of Poetiq. “They are simply not the best tools for deep reasoning. That’s why efforts to improve their problem-solving skills are so slow and expensive. For ARC-AGI 1 and 2, we used recursive self-improvement to produce specialized agents in a matter of hours. It demonstrates how much we can help with problems that have been too hard or too expensive for LLMs alone.”
At the same time, broader industry data highlights why Poetiq’s approach matters. An MIT study published in August 2025, which analyzed 300 public AI implementations, found that despite enterprises investing $30–$40 billion in generative AI, 95% of organizations are seeing zero return on investment. Consequently, AI use cases that fail to deliver ROI are prime candidates for Poetiq’s technology, as it enhances reasoning across both proprietary and in-house models.
AI Authority Trend: Ivo Raises $55 Million Series B to Turn Contracts Into a System of Record for Business Intelligence
Investors have also taken notice. “That Poetiq managed to top ARC-AGI within six months of launching is remarkable,” said Philipp Stauffer, General Partner at FYRFLY Venture Partners. “Rather than compete against frontier models, their team of six found a way to coax more intelligence from every LLM available. Poetiq will be a must-have for companies trying to make AI work for real-world business applications.”
Similarly, Gyan Kapur, co-Managing Partner at Surface Ventures, emphasized Poetiq’s flexibility. “Poetiq is one of the rare AI startups that doesn’t need to outcompete frontier models or pick sides,” he said. “It can enhance any combination of LLMs, any native AI platform, and any AI use case. Poetiq can provide better performance at lower costs across diverse use cases by sitting on top of foundation models, and that is a unique position to be in.”
Originally developed by AI researcher François Chollet, the ARC-AGI benchmark evaluates an AI system’s ability to generalize problem-solving in a human-like way. In early December, Poetiq set a new state of the art on the semi-private ARC-AGI-2 evaluation, surpassing Gemini 3 Deep Think at half the cost per task. Shortly thereafter, Poetiq integrated OpenAI’s GPT-5.2, achieving 75% accuracy on the public evaluation set, a 16-point jump over the previous benchmark leader. In response, OpenAI co-founder and President Greg Brockman tweeted that Poetiq is “exceeding the human baseline on ARC-AGI-2 with gpt-5.2.”
AI Authority Trend: ClickHouse Raises $400 Million Series D to Accelerate AI Data Infrastructure
To share your insights, please write to us at info@intentamplify.com
