Models
Enterprise
Subscribe
Resource
Documentation
Console
2 MIN READ

Models & Compatibility

Learn how to leverage our model ecosystem — explore supported LLMs, check feature compatibility, and understand parameter mapping.

Overview

Choose models by modality, quality, latency, and cost. Compatibility differs by provider — always verify parameters before scaling.

Decision checklist

  1. Input type: text / image / audio
  2. Required features: tools, JSON mode, vision, long context
  3. Budget: cost per 1K tokens and expected volume
  4. Latency: interactive UI vs batch jobs

Choosing a model

FAQ

How do I choose a model?

Start with a model that matches your input type. Run a small evaluation set, then scale only after quality and cost look acceptable.

Tags: Models

Can I switch models without changing my app?

Often yes — keep the request schema stable and change only the model field. Re-test tools / JSON mode when you switch families.

What is a good migration path?
text 
smoke test → shadow traffic → percentage rollout → full cutover

Parameters

Common fields

Field Notes
temperature Higher = more random
max_tokens Caps completion length
top_p Nucleus sampling

FAQ

What parameters are supported?

Different models support different parameters. Use each model's docs for supported fields and expected formats.

Tags: Parameters

Why was my parameter ignored?

Some gateways silently drop unsupported fields. Check the model card; prefer failing fast in your own validation layer.

Compatibility matrix (illustrative)

Feature Small text Flagship text Vision
Streaming Yes Yes Yes
Tools Limited Yes Varies
JSON mode Varies Yes Varies

Always confirm against the live model list — capabilities change as providers ship updates.