Inception

Diffusion language models

Tier

Moderate

Headquarters

Palo Alto, CA, USA

Founded

2024

Flagship

Mercury

Models tracked

1

Artificial Analysis

Top intelligence

32.8

Artificial Analysis Index

Cheapest model

$0.38

blended /1M

Fastest output

1234

tokens / sec

Flagship

Inception’s top-scoring model on the Artificial Analysis Intelligence Index.

Most intelligent model

Mercury 2

Released 2026-02-20

Intelligence

32.8

Coding

30.6

Math

In $/1M

$0.25

Out $/1M

$0.75

Speed t/s

1234

Benchmark breakdown

GPQA Diamond
77
Humanity’s Last Exam
16
SciCode
39
τ²-Bench (agentic)
71
Terminal-Bench Hard
27
IFBench
70

Models & benchmarks

Every Inception model on file — intelligence, coding, price, and speed.

ModelReleasedIntelCodingMathIn $/1MOut $/1MSpeed t/sLatency s
Mercury 22026-02-2032.830.6$0.25$0.7512343.57
Via Artificial Analysis

House strengths

Where Inception’s lineup over-indexes, averaged across its models.

Overview

A startup building diffusion-based language models (Mercury) aimed at very high generation speed.

Visit Inception