Mercury Decide

Mercury Decide is a decision model from Inception, the company that makes the Mercury family of diffusion language models. A System One model reads a state and answers typed questions about it, and Mercury Decide is Inception’s entry in that class, alongside Jev. OpenRouter describes it as a structured decision model served as a System One endpoint. You send a state and typed questions. It returns a choice, a score or a yes/no answer, each with a probability. OpenRouter says that probability is taken from the model rather than written out as text, so there is no generated prose to parse.

It is not OpenAI-compatible. According to AlphaSignal’s report on the launch, you call OpenRouter’s /v1/systemone endpoint instead of a chat completions route. OpenRouter lists the input as text and the output as decisions. The only endpoint is inception/mercury-decide-20260930:free, and Inception is the only provider. Inception makes diffusion LLMs, which generate text in parallel rather than one token at a time. The launch materials we found do not say how Mercury Decide is built beyond that, and no weights are published.

The primitives map to the three answer types: a choice picks one option from a list, a score returns a level, and a yes/no answer is what the site calls Noul. Each comes with a probability. The limits on options, levels and questions per call are not published as of 2026-10-01.

Inception’s launch post calls Mercury Decide “the most intelligent decision model on @OpenRouter (JevBench v1.4)” and says it makes “up to 14 decisions per second”. Both numbers are reported by Inception and were not independently run. The model does not appear on Benchmark Heaven’s public JevBench board as of 2026-10-01, and the launch post does not link to the evaluation behind the claim. Treat the ranking as a vendor statement until someone reproduces it.

OpenRouter says it is built for triage, routing, moderation, agent step selection and evaluation loops. These are calls where the answer comes from a fixed set and you want a probability to gate on, for example sending a ticket to a queue, picking the next step for an agent, or scoring an output against a rubric.

It does not write text, so anything that needs an explanation or a draft belongs with a language model. With no published accuracy figure outside Inception’s own claim and no calibration data, test the probabilities on your own cases before you act on a threshold.

The model is free on OpenRouter as inception/mercury-decide:free during early access, with a 32,768-token context. Inception has published no paid price, rate limit or SDK as of 2026-10-01.

Question types Choice Score Noul Max Choice optionsNot documented Score levelsNot documented Questions per callNot documented Total context32,768 tokens State budgetNot documented Rate limitNot published as of 2026-10-01. EndpointPOST /v1/systemone on OpenRouter SDKs OpenRouter lists a 32,768-token context and a 29,491-token maximum completion for the endpoint. The limits on options per Choice, levels per Score and questions per call are not published as of 2026-10-01.

inception/mercury-decide-20260930:free, 30 Sep 2026, The only endpoint OpenRouter lists, served by Inception alone. OpenRouter lists the model as inception/mercury-decide:free. Release notes

What people use Mercury Decide for, one page per pattern. Workflow controlStarter

Send a support ticket to Jev once with every question attached. Category comes back as a selected label, severity and frustration as numbers on scales you wrote, refund intent as a probability. Your code reads those values and decides what happens to the ticket. Choice Score Noul Workflow controlIntermediate

One Jev call reads an incoming request and returns its intent as a label plus a difficulty rating on a scale you wrote. Your router reads both numbers and picks the handler: deterministic code, a cheap model, an expensive one, or a human queue. Choice Score Safety and qualityIntermediate