KeyveraKeyvera
Get API key ↗
PUBLIC MODEL PLAZA / LIVE CATALOG

Choose the model.
Keep the API.

Explore the model routes currently published through Keyvera. One OpenAI-compatible integration lets your application switch capability, speed, and cost by changing the model identifier.

Get your API key ↗See live pricing →
MODEL PLAZA CONNECTED54PUBLISHED ROUTES
12ACCESS GROUPS
54 MODELS AVAILABLE54 PUBLISHED ROUTES12 ACCESS GROUPSONE API FORMATPUBLISHED TOKEN RATESMODEL-ID ROUTING54 MODELS AVAILABLE54 PUBLISHED ROUTES12 ACCESS GROUPSONE API FORMAT
01 / ROUTE DIRECTORY

Published model routes.

Routes may appear in more than one access group. Availability and model IDs are sourced from the public Keyvera Model Plaza.

Exact pricing ↗
54ROUTES
MATCHED
OpenAI-compatible

Deepseek V4 Flash

deepseek-v4-flash
Anthropic-compatible

Claude Haiku 4 5

claude-haiku-4-5
Anthropic-compatible

Claude Opus 4 6

claude-opus-4-6
Anthropic-compatible

Claude Opus 4 7

claude-opus-4-7
Anthropic-compatible

Claude Opus 4 8

claude-opus-4-8
Anthropic-compatible

Claude Sonnet 4 6

claude-sonnet-4-6
Anthropic-compatible

Claude Sonnet 5

claude-sonnet-5

Catalog note: The Model Plaza is the operational source of truth for currently published routes. The broader Keyvera platform supports 50+ available models; route exposure can vary by access group and availability.

MODEL ROUTING GUIDE

How to choose AI models through Keyvera.

The models page is the decision layer for developers comparing capability, cost, and integration effort. It explains how Keyvera exposes many model routes through one OpenAI-compatible API instead of forcing teams to maintain separate provider clients.

For SEO and practical evaluation, this page intentionally names the model-routing concepts buyers search for: AI API gateway, OpenAI-compatible endpoint, model catalog, token pricing, provider routing, published model IDs, access groups, and production LLM infrastructure.

Model IDs are the switch

With Keyvera, the model identifier is the main routing control. Your application can keep the same request body, streaming behavior, authentication pattern, and response parsing while changing the upstream model selected for a task.

This is useful for products that need separate routes for coding, support automation, content analysis, extraction, classification, long-context review, and reasoning-heavy workflows.

Provider choice stays flexible

The catalog includes routes across OpenAI, Anthropic, xAI, DeepSeek, Kimi, Zhipu, and MiniMax. A team can evaluate provider strengths without turning every evaluation into a new SDK integration, a new billing workflow, and a new credential rotation problem.

Provider pages link from this directory so search engines and users can understand each model family in context.

Operational signals matter

Availability can differ between the platform, an upstream provider, and a specific model route. The public page shows the catalog shape, while authenticated dashboard views can expose more detailed route information for active users.

Before production rollout, teams should compare model quality, latency, token cost, cache behavior, and fallback requirements.

Where to go next

Use Pricing for exact published prices and token-rate explanation, Integrations for setup guidance, and Security for data-handling boundaries.

The broader Keyvera platform supports 50+ available models; the live public route list reflects what is currently published by access group.

ONE FORMAT / MANY MODELS

Switch the model ID.
Keep shipping.

One endpoint and one prepaid balance across the Keyvera catalog.

Compare live pricing ↗
VIEW ROUTE STATUS ↗