Skip to main content
Back to Explore

Inception: Mercury 2

Inception·Released 2026-03-04
128K ctx

About

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...

Pricing

Input

$0.25

per 1M tokens

Output

$0.75

per 1M tokens

Blended

$0.38

per 1M tokens

Cheaper than 57% of models. Median price is $0.54/1M tokens.

Cost Calculator

Tokens per day1M
100K100M

Daily

$0.38

Monthly

$11.25

vs. Similar Models

DeepSeek V3.2 Exp (Reasoning)Q:+0.1
$0.31-17%
NVIDIA Nemotron 3 Super 120B A12B (Reasoning)Q:+0.1
$0.41+10%
Claude 4 Opus (Non-reasoning)Q:+0.2
$30.00+7900%
Claude 4 Sonnet (Non-reasoning)Q:+0.2
$6.00+1500%

Performance

1054

tokens/sec

Faster than 100% of models

3.25

seconds

Faster than 19% of models

3.25

seconds

Faster than 49% of models

Market Median

94 tok/s

1016% faster

Median TTFT

1.11s

192% slower

Throughput/Dollar

2810

tok/s per $/1M

Speed Comparison

LFM2.5-1.2B-Instruct
537 tok/s-49%
LFM2 1.2B
532 tok/s-50%
LFM2.5-VL-1.6B
508 tok/s-52%

Context Window

128K

tokens

Larger than 16% of models

Max Output

50K

tokens

39% of context

Benchmarks

MMLU-ProNot evaluated
GPQA Diamond
77.0%
HLE
15.5%
LiveCodeBenchNot evaluated
SciCode
38.7%
TerminalBench Hard
26.5%
MATH-500Not evaluated
AIMENot evaluated
AIME 2025Not evaluated
IFBench
69.8%
Long Context Recall
36.3%
Tau2
70.8%
Market AverageTop Score

Quick Compare

Similar Models

Compare all 7 models