Skip navigation
The right model depends on the task.
0:00 / 0:23

The right model depends on the task.

NVIDIA

2.7M followers

Visit on X
327 likesDownloadOriginal post

97,320 views

The right model depends on the task. NVIDIA NeMo Switchyard helps developers route each agent workflow step across a chosen model pool based on their own quality, latency and cost criteria. Kari Briski joins MTS to explain why agent workflows need model routing.

33 Comments

Comments from the original post on X (Twitter)

  • NVIDIA's profile picture

    @nvidia

    Learn more:

    37 likes

  • Brent Liang's profile picture

    @liangsays

    @MTSlive so awesome to have kari on the show!

    5 likes

  • Denys Khomyn's profile picture

    @denys_khomyn

    @ChrisJBakke @MTSlive NVIDIA monitoring the situation 🤘

    4 likes

  • DataVLT's profile picture

    @DataVLTAI

    @MTSlive The smarter the routing, the more important the data behind each decision becomes. Model choice is only one part of the stack.

    2 likes

  • PixelSilicon's profile picture

    @pixelsilicon

    @MTSlive Model routing by task is becoming essential once agent workflows get multi step. One model for everything is rarely optimal on cost or latency.

    1 like

  • KVCacheStore LLC's profile picture

    @KVCacheStore

    @MTSlive I'm a contributor to Switchyard. If anyone wants help setting it up let me know.

    1 like

  • Joel Ramirez's profile picture

    @Joelramirez1968

    @MTSlive ⚓️✅❌🚀🎇

    1 like

  • Glitch Truth's profile picture

    @glitchtruth

    @MTSlive Doesn't matter which model wins the routing fight, every single path still ends at an Nvidia GPU.

    1 like

  • OGNVDASOL's profile picture

    @OGNVDASOL

    @MTSlive @elonmusk

    1 like

  • Zee B-side's profile picture

    @ZeeBside

    @MTSlive oxi, now proibida no mundo todo

  • Ce Ce | AI & Systems Infra's profile picture

    @pusulainfra

    @MTSlive model routing based on latency + cost is the part most teams still underestimate

  • cameron_h23's profile picture

    @cameron_h23

    @MTSlive Model routing is becoming core infrastructure for agent systems—use the expensive reasoning model where it matters, optimize everything else around it.

  • AI Mastery Guide's profile picture

    @aiseomastery

    @MTSlive Smart routing saves real money

  • Robert Hero's profile picture

    @robertahero

    @MTSlive The right model depends on the task.

  • Marissa's profile picture

    @marissarobinart

    @MTSlive 👍

  • J. LEE's profile picture

    @JWPAX3051

    @MTSlive Love how NeMo Switchyard lets us pick the perfect model for each step—smart, flexible, and cost‑savvy!

  • MoM's profile picture

    @mixtureofmodels

    @MTSlive

  • Abhi's profile picture

    @SpendLensAI

    @MTSlive Routing only works after teams save representative workflows and set a quality floor.

  • Christopher Dean's profile picture

    @coderchrisdean

    @MTSlive Per-step routing is right, and it only works if you have per-step evals: without them you're routing on vibes and the cheap model quietly tanks one node of the workflow.

  • nock's profile picture

    @nockrail

    @MTSlive rubin vs tpu pileup this month is the whole story. everyone shipping silicon at once.