# Inception

> Inception builds diffusion-based large language models (dLLMs) that generate text in parallel instead of one token at a time. Its flagship Mercury family is positioned as the first commercial-scale diffusion LLM, reporting 5–10x faster throughput than traditional autoregressive models at competitive quality, with Mercury Coder targeting code generation and Mercury 2 reasoning workloads.

- **Founded:** 2024
- **Headquarters:** Palo Alto, California, United States
- **Founders:** Stefano Ermon (Co-founder & CEO), Aditya Grover (Co-founder), Volodymyr Kuleshov (Co-founder)
- **Team size:** 37
- **Products:** Mercury, Mercury Coder, Mercury 2, Mercury Edit 2, Inception Platform / API
- **Notable:** Launched Mercury Coder on February 26, 2025 as the first commercial-grade diffusion LLM., Reported throughput above 1,000 tokens per second on a single NVIDIA H100., Closed a $50M seed round in November 2025 led by Menlo Ventures.

## Products & services

- **Mercury** — Family of commercial-scale diffusion LLMs that denoise blocks of tokens in parallel rather than generating one token at a time.
- **Mercury Coder** — Diffusion LLM tuned for code generation; OpenAI API compatible, priced at $0.25/M input and $0.75/M output tokens.
- **Mercury 2** — Next-generation reasoning dLLM aimed at complex applications and real-time agent workflows.
- **Mercury Edit 2** — Compact, coding-focused diffusion model designed for latency-sensitive code-edit loops.
- **Inception Platform / API** — Hosted API and on-prem deployment options for enterprise access to Mercury models.

## Achievements

- Launched Mercury Coder on February 26, 2025 as the first commercial-grade diffusion LLM.
- Reported throughput above 1,000 tokens per second on a single NVIDIA H100.
- Closed a $50M seed round in November 2025 led by Menlo Ventures.
- Backed by Andrew Ng and Andrej Karpathy as angel investors.
- Published the Mercury technical paper on arXiv (2506.17298).

## Latest updates

- **2025-11** — Announced $50M seed round led by Menlo Ventures.
- **2025-11** — Launched Mercury 2 and Mercury Edit 2.
- **2025-06** — Published Mercury technical paper on arXiv.
- **2025-02** — Unveiled Mercury Coder as the first commercial-scale diffusion LLM.

## Links

- Website: https://www.inceptionlabs.ai/
- LinkedIn: https://www.linkedin.com/company/inception-labs-ai/
- Twitter/X: https://x.com/_inception_ai

---

Profile page: https://yespress.io/inception
Published by YesPress — https://yespress.io
Last updated: 2026-06-29