Inception Unveils Mercury 3, Claiming up to 8X Faster AI Generation

WHO:

MarkTechPost - Inception released Mercury 3, a diffusion large language model (LLM).

Techmeme - Jason Hiner of The Deep View reported on Inception’s Mercury 3 launch.

MarkTechPost - Inception’s benchmark table compared Mercury 3 with GPT-6 Luna and Claude Haiku 5.5.

WHAT:

MarkTechPost - Inception says Mercury 3 runs at 1,068 tokens per second and supports a 1M-token context.

MarkTechPost - Inception’s table shows Mercury 3 leading GPT-6 Luna and Claude Haiku 5.5 on 8 of 12 benchmarks.

Techmeme - Inception describes Mercury 3 as “comparable to Luna and Haiku models” and claims it is 4x to 8x faster, generating approximately 1,000 tokens per second.

WHY:

Techmeme - The article places Mercury 3’s launch in the context of an AI industry that has invested heavily in LLMs and the large data centers needed to scale them.

HOW:

MarkTechPost - Mercury 3 is described as a diffusion LLM.

Techmeme - Mercury 3 is described as diffusion-based.

(UPDATED October 21, 2026 01:30 UTC)

Inception has released Mercury 3, a diffusion-based large language model it says can generate 1,068 tokens per second and handle a 1-million-token context. Inception says Mercury 3 is four to eight times faster than comparable models, and its benchmark table shows it ahead of GPT-6 Luna and Claude Haiku 5.5 on 8 of 12 tests. The launch comes as the AI industry continues investing heavily in large language models and the data centers needed to scale them.