# Pokee AI

> Pokee AI is a Bellevue-based startup building foundation AI agents that plan, reason, and act across thousands of internet tools. Founded in 2024 by former Meta applied-reinforcement-learning lead Zheqing (Bill) Zhu, the company uses reinforcement learning rather than pure LLM function-calling to sequence tool use reliably, claiming its foundation model outperforms GPT-4o, Claude 3.7, and Gemini 2.5 Pro on tool selection. Pokee raised a $12M seed round in 2025 led by Point72 Ventures, ships an enterprise platform with private/on-prem deployment, and open-sourced PokeeResearch-7B, a state-of-the-art small deep-research agent.

- **Founded:** 2024
- **Headquarters:** Bellevue, Washington, United States
- **Founders:** Zheqing (Bill) Zhu (Founder & CEO)
- **Team size:** ~18 employees (2025)
- **Products:** Pokee AI Agent Platform, Pokee Foundation Tool-Use Model, PokeeResearch-7B
- **Notable:** Raised $12M seed round in 2025 led by Point72 Ventures with Qualcomm and Samsung participating., Built a foundation tool-use model the company says beats GPT-4o, Claude 3.7, and Gemini 2.5 Pro on function calling., Released PokeeResearch-7B, state-of-the-art among open 7B deep-research agents, leading 7 of 10 benchmarks.

## Products & services

- **Pokee AI Agent Platform** — An enterprise AI-agent platform that executes multi-step workflows across 90+ SaaS tools (covering 1,000+ integrations) from natural-language instructions, with private cloud, on-premise, or on-device deployment and end-to-end encryption.
- **Pokee Foundation Tool-Use Model** — A proprietary reinforcement-learning-trained foundation model for tool usage, planning, and reasoning that the company says surpasses GPT-4o, Claude 3.7, and Gemini 2.5 Pro on function calling and extends to 6,000+ tools with high tool-selection accuracy.
- **PokeeResearch-7B** — An open-source 7-billion-parameter deep-research agent trained with Reinforcement Learning from AI Feedback (RLAIF) and a robust reasoning scaffold; reasons across the web, retrieves evidence, and self-verifies. Weights and inference code are released under Apache 2.0 on Hugging Face and GitHub.

## Achievements

- Raised $12M seed round in 2025 led by Point72 Ventures with Qualcomm and Samsung participating.
- Built a foundation tool-use model the company says beats GPT-4o, Claude 3.7, and Gemini 2.5 Pro on function calling.
- Released PokeeResearch-7B, state-of-the-art among open 7B deep-research agents, leading 7 of 10 benchmarks.
- Reached public beta with enterprise design partners within roughly a year of founding.
- Filed provisional patents on long-context model design.

## Latest updates

- **2025-07** — Announced $12M seed round led by Point72 Ventures to build AI agents that automate online workflows.
- **2025-10** — Open-sourced PokeeResearch-7B, a SOTA 7B deep-research agent, with weights and inference code on Hugging Face and GitHub.

## Links

- Website: https://pokee.ai
- LinkedIn: https://www.linkedin.com/company/pokee-ai
- Twitter/X: https://x.com/Pokee_AI
- GitHub: https://github.com/Pokee-AI
- YouTube: https://www.youtube.com/@pokeeai

---

Profile page: https://yespress.io/pokee-ai
Published by YesPress — https://yespress.io
Last updated: 2026-07-03