What is Inception Mercury?

Inception Mercury is a Developer tools tool: Diffusion LLM API built for very fast, low-latency text generation.

How much does Inception Mercury cost?

Inception Mercury uses a Freemium model; see the vendor's pricing page for current plans.

What are the best alternatives to Inception Mercury?

Community-listed alternatives to Inception Mercury include Cursor, Exa, RunInfra, Auth0.

What does Inception Mercury integrate with?

Inception Mercury integrates with OpenAI SDK, OpenRouter, Baseten, Vercel, LiteLLM, LangChain and more.

Which platforms does Inception Mercury support?

Inception Mercury is available on web.

Inception Mercury

Diffusion LLM API built for very fast, low-latency text generation

FreemiumAI assistants Developer tools

Visit website

Inception Mercury is a developer tools tool for Developers and Enterprise, offered as freemium. Popular alternatives include Cursor, Exa, RunInfra.

About Inception Mercury

Inception Mercury is the API platform for Mercury, a family of diffusion large language models (dLLMs) from Inception, a startup led by Stanford professor and CEO Stefano Ermon that raised a $50 million seed round led by Menlo Ventures in November 2025. Diffusion models refine whole blocks of text in parallel instead of writing one token at a time, which Inception sells as fast, low-latency generation.

The flagship Mercury 3, released October 19, 2026, has a 1M-token context window, adjustable reasoning effort and tool calling, and Inception measured it at 1,068 tokens per second on NVIDIA GPUs. Mercury Voice (enterprise only) targets voice agents with time-to-first-token under 170 ms, and Mercury Decide returns calibrated probabilities for classification. The API is OpenAI-compatible and also served through OpenRouter, Baseten and Vercel. New keys get 100 million free tokens; Mercury 3 then lists at $0.20 per million input and $0.70 per million output tokens, cut 50% at launch, and Enterprise adds custom limits and SLAs.

It fits latency-sensitive coding, voice and web agents. DeepSeek is a cheap autoregressive option with open weights, while Claude and Gemini offer wider multimodal model ranges. Inception says Mercury 3 outperforms Claude Haiku 5.5, but it has not published the benchmark scores, and Mercury 2 and 2.5 are already closed to new users.

Inception launched Mercury 3 on Product Hunt on October 22, 2026 and finished #11 of the day. Buying tip: test your prompts with the free tokens and budget at list price, since the launch discount has no stated end date.

Key features of Inception Mercury

Mercury 3 diffusion LLM with a 1M-token context window
Measured 1,068 tokens per second output on NVIDIA GPUs
Adjustable reasoning effort to trade speed against thoroughness
Tool calling with parallel and chained calls for agents
Mercury Voice with under 170 ms time-to-first-token for voice agents
Mercury Decide classification model with calibrated probabilities
OpenAI-compatible API plus Python and TypeScript SDKs
Also available through OpenRouter, Baseten and Vercel

Pros and cons

Pros

Very high output speed suits real-time and agent loops
Low list price of $0.20 input and $0.70 output per million tokens
100 million free tokens with no card to start
Drop-in for OpenAI SDK, LiteLLM and LangChain stacks
Cached input billed 90% below standard input

Cons

Quality claims against rivals come without published benchmark scores
Text-focused lineup with no image or video generation
Launch discount has no stated end date, so costs may double
Fast model turnover: Mercury 2 and 2.5 already closed to new users
Mercury Voice is open only to enterprise customers through sales

Inception Mercury pricing

Inception Mercury uses a freemium model.

Plan	Price	Includes
Free	$0	100M free tokens for new API keys, all models, no card
Mercury 3	$0.20 in / $0.70 out / 1M tokens	Launch price 50% off: $0.10 in, $0.35 out; cached input $0.02
Mercury Voice	$0.40 in / $1.50 out / 1M tokens	Launch price 50% off: $0.20 in, $0.75 out; enterprise access via sales
Mercury Decide	$0.04 in / 1M tokens	Launch price 50% off: $0.02 in; output tokens free
Enterprise	Custom	Custom rate limits, SLA, volume pricing, dedicated capacity

Source: www.inceptionlabs.ai · checked on Oct 25, 2026

Are you the maker of Inception Mercury?

Show a price-verified badge on your site. It displays when we last checked your prices on your own pricing page and links back to this page.

Embed code

<a href="https://happyvalley.dev/tools/inception-mercury"><img src="https://happyvalley.dev/api/badge/inception-mercury?v=price" alt="Prices verified on Happy Valley" height="32"></a>

How Inception Mercury compares in Developer tools

Rank by votes: #2 / 57
Starting price: $0
Free alternatives: 12

Inception Mercury ranks #2 of 57 developer tools tools by community votes and is not directly comparable on price.

Common use cases

1Powering code completion and coding agents where latency matters
2Running voice agents that must answer within a fraction of a second
3Web and browser agents that make many sequential model calls
4High-volume classification and routing with Mercury Decide
5Swapping a slower model in an existing OpenAI-based app

Best for

Developers building latency-sensitive agents, voice bots or coding tools who want a fast, cheap text model behind an OpenAI-compatible API.

Not ideal for

Teams that need multimodal models, open weights or the widest ecosystem of hosted tools (Gemini, DeepSeek, Claude).

Integrations

OpenAI SDK
OpenRouter
Baseten
Vercel
LiteLLM
LangChain
AISuite
Kilo Code

Our take

Inception Mercury is worth testing whenever response speed is the bottleneck, and the free token allowance makes that cheap to do. Benchmark it on your own tasks before committing, because the published comparisons are claims rather than scores.

FAQ

What is Inception Mercury?: Inception Mercury is a Developer tools tool: Diffusion LLM API built for very fast, low-latency text generation.
How much does Inception Mercury cost?: Inception Mercury uses a Freemium model; see the vendor's pricing page for current plans.
What are the best alternatives to Inception Mercury?: Community-listed alternatives to Inception Mercury include Cursor, Exa, RunInfra, Auth0.
What does Inception Mercury integrate with?: Inception Mercury integrates with OpenAI SDK, OpenRouter, Baseten, Vercel, LiteLLM, LangChain and more.
Which platforms does Inception Mercury support?: Inception Mercury is available on web.
Who is Inception Mercury for?: Developers building latency-sensitive agents, voice bots or coding tools who want a fast, cheap text model behind an OpenAI-compatible API.

Alternatives to Inception Mercury

See all 10

Cursor
▲ 0
AI-first code editor with agents, now owned by SpaceX
Freemium · $8Developers
checked Oct 20
Exa
▲ 0
Web search and content API built for AI agents and LLM apps
FreemiumDevelopers
checked Oct 22
RunInfra
▲ 0
Open-model API with flat monthly plans for coding agents
Paid · $10Developers
checked Oct 22
Auth0
▲ 0
Identity platform for applications
Freemium · $35Developers
checked Oct 20
Postman
▲ 0
API platform for testing, mocking, documenting and sharing APIs
Freemium · $9Developers
checked Oct 20
Apidog
▲ 0
All-in-one API design, testing, mocking and docs; a Postman rival
Freemium · $12Developers
checked Oct 20

More Developer tools tools

See all 57

Developer Tools
▲ 2
173 free web, DNS, SSL, SEO & developer tools — fast, live, no sign-up
FreeFreelancers
Windmill
▲ 0
Open-source developer platform for scripts, flows and apps
Open source · $10Developers
checked Oct 20
Sentry
▲ 0
Error and performance monitoring
Freemium · $29Developers
checked Oct 20
Whiteboard
▲ 0
Open-source desktop canvas for reviewing coding agent work
Open sourceDevelopers
checked Oct 25
Snyk
▲ 0
Developer security for code and dependencies
Freemium · $25Developers
checked Oct 20
Zalcro
▲ 0
Plans an app's architecture before you prompt an AI builder
Freemium · $19Developers
checked Oct 24

Reviews

No reviews yet. Be the first to share real experience.