NexoIP AI Tools hub
Homepage
ES EN
LLM routing layer

Start here if you still need to choose the right comparison layer.

This page is intentionally lighter than the matrix. It exists to send you to the right LLM decision layer before the tables start repeating the same argument.

ℹ️

Boundary rule

This page is the routing layer. It should stay lighter than the matrix, less vendor-specific than Provider Compare, less opinionated by task than Model Fit Radar, and less operational than Workflow Recipes. Hardware and self-host capacity decisions now live in Inference Hardware Guide.

Decision layers

4

Vendor, model, scenario and workflow

Models in matrix

10

Raw technical rows

Providers

5

Vendor posture lanes

Recipes live

6

Practical operating flows

How to move through the LLM stack

  1. 1

    Vendor

    Open Provider Compare if the question is still managed API vs multimodal vs open-weight

    This is the vendor layer. Stay there until the operating posture is narrow enough.

  2. 2

    Model

    Open the matrix when the decision is already row-level

    Use the matrix for context, price, deployment and open-weight details.

  3. 3

    Scenario

    Open Model Fit Radar when you need a quick lane by task

    Use it for coding, multimodal, long reasoning, cost-sensitive and local-open picks.

  4. 4

    Operation

    Open Workflow Recipes when the question becomes how to run the lane

    This is the practical layer for review loops, retrieval, browser workers and local-first flows.

LLM technical matrix

Next layer

Row-level comparison for context, spend, deployment and open-weight posture.

Open matrix

Provider compare

Next layer

Vendor-level comparison for managed API, multimodal, self-host and price posture.

Open providers

Model fit radar

Next layer

Scenario-first picks when the question is which model lane fits the task.

Open radar

Workflow recipes

Next layer

Operational playbooks for coding review, retrieval, browser flows and local-first execution.

Open recipes

Which lane should you open next?

Layer Primary question Use when Avoid when
Provider compare Which vendor posture fits the problem? You are still choosing between managed API, multimodal, self-host or router-friendly lanes. The choice is already between concrete model rows.
LLM technical matrix What do the rows look like technically? You need context, spend, deployment and open-weight details before testing. You still need a scenario-first or workflow-first recommendation.
Model fit radar Which model lane fits this task fastest? You need quick picks for coding, multimodal, long reasoning or cheap routing. You need raw vendor or row-level detail first.
Workflow recipes How should this lane run in practice? Provider and model choice are narrow enough and the next problem is execution. You still do not know what to buy or compare.