AI Tools & Platforms API & Backend Developer ToolsPaid

Mercury

Mercury is Inception's family of diffusion large language models, served through an OpenAI-compatible API. Mercury 3, launched in October 2026, is the company's flagship reasoning model with a 1M-token context.

0.0 (0)

Visit Website

Founder

Inception

Active

PlanMercury 3 - $0.20 input / $0.70 output per 1M tokens (launch: 50% off, $0.10 / $0.35); cached input 90% off · Mercury Voice - $0.20 input / $0.75 output per 1M tokens (launch-discount price shown on site) · Enterprise - contact sales - dedicated capacity, fine-tuning, private deployment, custom SLAs

BusinessB2B

About Mercury

What is Mercury?

Mercury is Inception's family of diffusion LLMs. Instead of generating one token at a time, the models refine output in parallel, which Inception says makes them faster and cheaper than autoregressive LLMs.

Models

Mercury 3 (Oct 2026): flagship reasoning model, 1M-token context; the company claims over 1,000 tokens/s on NVIDIA GPUs.
Mercury Voice: for voice agents, under 170 ms time to first token (vendor figure).
Mercury Edit 2: next-edit and fill-in-the-middle code completion.
Mercury Decide: calibrated decisions for classification, routing and scoring.

API

OpenAI-compatible API with chat, FIM, edit and decisions endpoints. Also offered via AWS Bedrock, Azure Foundry, Baseten, OpenRouter and Vercel; integrations with Cursor, Zed, VS Code, LangChain and Vapi.

Pricing

Mercury 3: $0.20 / $0.70 per 1M input/output tokens (50% off at launch), cached input 90% off.

Company

Inception, Redwood City, California. $50M round led by Menlo Ventures (Nov 2025).

Value & Audience

Value Proposition

Gives developers LLMs with very low latency and per-token cost through a drop-in OpenAI-compatible API, aimed at real-time uses such as voice agents, code completion and high-volume pipelines.

Problem Solved

Autoregressive LLMs generate tokens sequentially, which limits speed and raises the cost of latency-sensitive and high-volume applications.

Target Audience

Developers Data Scientists & ML Engineers Startups Enterprises Researchers

Use Cases

Share this product

Product founder?

Get Badge

Tech Stack

Other

Framer

Social & Links

Social

Twitter LinkedIn Discord

Additional Links

docs.inceptionlabs.ai platform.inceptionlabs.ai inceptionlabs.ai inceptionlabs.ai inceptionlabs.ai chat.inceptionlabs.ai

Loading ratings...

Loading comments...

Gallery

Feedback & Roadmap

Loading feedback…

Mercury FAQ

Common questions about Mercury's features, pricing, and use cases

What does Mercury do?

Diffusion LLMs from Inception for fast, low-cost inference Gives developers LLMs with very low latency and per-token cost through a drop-in OpenAI-compatible API, aimed at real-time uses such as voice agents, code completion and high-volume pipelines.

Is Mercury free to use?

Mercury is a paid product, starting at Mercury 3 - $0.20 input / $0.70 output per 1M tokens (launch: 50% off, $0.10 / $0.35); cached input 90% off. Full pricing details are listed on the Mercury product page on TechLogHub.

What problem does Mercury solve?

Autoregressive LLMs generate tokens sequentially, which limits speed and raises the cost of latency-sensitive and high-volume applications.

Who is Mercury built for?

Mercury is built for Developers, Data Scientists & ML Engineers, Startups, Enterprises, Researchers.

What tech stack does Mercury use?

Mercury is built with Framer.

What category is Mercury listed under?

Mercury is listed in the AI Tools & Platforms category on TechLogHub, alongside curated alternatives with community ratings and pricing comparisons.

Mercury appears in

PricingMercuryplans, free tier & alternatives CategoryBest AI Tools & Platforms tools