modelgrep
I

Inception: Mercury 2

inception/mercury-2

69th smartest of 178ReasoningToolsJSON
Use via OpenRouter ↗
Intelligence
32.8
69th of 178
Design Elo
Speed
386
8th fastest
Latency
254ms
first token
Input price
$0.250
110th cheapest
Context
128K
50K max out

How it compares

Smarter than61%
of all ranked models
Faster than97%
of all ranked models
Cheaper than63%
of all ranked models

Overview

Mercury 2 is an extremely fast reasoning LLM, and the first reasoning diffusion LLM (dLLM). Instead of generating tokens sequentially, Mercury 2 produces and refines multiple tokens in parallel, achieving...

Benchmarks

independent · via OpenRouter
Artificial Analysis59th percentile
Intelligence Index
32.8
Coding Index
30.6
Agentic Index
39.7
GPQA Diamond
77%
Humanity's Last Exam
16%
SciCode
39%
Tau²-Bench (agentic)
71%

Providers & pricing (1)

ProviderIn $/MOut $/MUptime
Inception$0.250$0.75099.8%

Specifications

Context window128K
Max output50K
Knowledge cutoff
Input modalitiestext
Output modalitiestext
Prompt caching
Cache read price$0.025/M
ModeratedNo

Mercury 2 FAQ

How much does Mercury 2 cost?

Mercury 2 costs $0.250 per million input tokens and $0.750 per million output tokens via OpenRouter, making it 110th cheapest of 298 paid models.

How smart is Mercury 2?

Mercury 2 scores 32.8 on the Artificial Analysis Intelligence Index, ranking 69th of 178 benchmarked models, with a GPQA Diamond score of 77%.

How fast is Mercury 2?

Mercury 2 generates around 386 tokens per second with 254ms time-to-first-token (p50), the 8th fastest tracked model.

What is Mercury 2's context window?

Mercury 2 supports a 128K-token context window and can output up to 50K tokens. It accepts text input.

Compare head-to-head