Inception Labs says its diffusion LLM is 10x faster than Claude, ChatGPT, Gemini

Inception Labs says its diffusion LLM is 10x faster than Claude, ChatGPT, Gemini

On a recent episode of the The New Stack Agents, Inception Labs CEO Stefano Ermon introduced Mercury 2, a large language model built on diffusion rather than the standard autoregressive approach. Traditional LLMs generate text token by token from left to right, which Ermon describes as “fancy autocomplete.” In contrast, diffusion models begin with a rough draft and refine it in parallel, similar to image systems like Stable Diffusion.

This parallel process allows Mercury 2 to produce over 1,000 tokens per second—five to ten times faster than optimized models from labs such as OpenAI, Anthropic, and Google, according to company tests. Ermon argues diffusion models better leverage GPUs, with support from investor Nvidia to optimize performance.

While Mercury 2 matches mid-tier models like Claude Haiku and Google Flash rather than top systems such as Claude Opus or GPT-4, Ermon believes diffusion’s speed and economic advantages will become increasingly compelling as AI applications scale.

Learn more from The New Stack about the latest developments around around large language model built on diffusion:

How Diffusion-Based LLM AI Speeds Up Reasoning

Get Ready for Faster Text Generation With Diffusion LLMs

Join our community of newsletter subscribers to stay on top of the news and at the top of your game.

Denne episoden er hentet fra en åpen RSS-feed og er ikke publisert av Podme. Den kan derfor inneholde annonser.

Episoder(300)

Bit Cloud’s next chapter starts after the AI builds your app

Bit Cloud’s next chapter starts after the AI builds your app

For many developers, turning an AI-generated prototype into maintainable software requires more than generating code—it requires infrastructure, collaboration, testing and review. In this episode of T...

1 Okt 31min

CloudBees just committed to an AI-first pivot. Here's why it matters for enterprise DevOps teams

CloudBees just committed to an AI-first pivot. Here's why it matters for enterprise DevOps teams

CloudBees CEO Mo Plassnig is leading the CI/CD company through a major transformation as generative AI reshapes software development. Returning to CloudBees eight years after joining through its acqui...

30 Sep 23min

A third option is emerging in the fight over AI and your data

A third option is emerging in the fight over AI and your data

The AI industry has faced a growing enterprise dilemma: companies want access to powerful proprietary AI models without risking sensitive data or intellectual property, while AI labs want to protect t...

23 Sep 30min

Drowning in AI pull requests: Harness's field CTO on code review and a Git repo built for agents

Drowning in AI pull requests: Harness's field CTO on code review and a Git repo built for agents

Harness Field CTO Martin Reynolds joins The New Stack to talk about what happens after coding agents start opening pull requests faster than anyone can review them. He explains how he first saw the bo...

7 Sep 26min

How to find failures without drowning in tracing data

How to find failures without drowning in tracing data

Traces provide a detailed view of a request’s journey through data, microservices and applications, helping SREs pinpoint where failures occur and resolve issues faster. But while tracing can reduce d...

3 Sep 31min

Why CPUs still matter in the age of AI agents

Why CPUs still matter in the age of AI agents

As AI evolves from conversational chatbots to autonomous agents, CPUs are becoming an increasingly important part of the infrastructure equation. In this episode, The New Stack speaks with Bhumik Pate...

11 Aug 26min

Why Doist Says Less AI Can Deliver More

Why Doist Says Less AI Can Deliver More

Doist CTO Gonçalo Silva says AI is reshaping software development, but success depends on restraint rather than rapid feature expansion. Instead of chasing every AI capability, Doist prioritizes “subt...

31 Jul 35min

Why your company should (try to) build its own AI SRE

Why your company should (try to) build its own AI SRE

As AI coding agents accelerate software development, they also create new challenges for site reliability engineers (SREs), who are increasingly responsible for debugging systems that no single human ...

30 Jul 30min

Populært innen Politikk og nyheter

giver-og-gjengen-vg
aftenpodden
forklart
aftenpodden-usa
popradet
fotballpodden-2
stopp-verden
dine-penger-pengeradet
rss-espen-lee-usensurert
det-store-bildet
rss-gukild-johaug
nokon-ma-ga
hanna-de-heldige
aftenbla-bla
rss-penger-polser-og-politikk
e24-podden
rss-ness
frokostshowet-pa-p5
bt-dokumentar-2
saken