# What is LLM failover?

Markdown version of https://kymaapi.com/glossary/llm-failover, for agents and crawlers. Same content as the HTML page.

- Generated: 2026-10-28T21:43:30.489Z
- HTML page: https://kymaapi.com/glossary/llm-failover
- Model catalog: https://kymaapi.com/models.md
- Pricing: https://kymaapi.com/pricing.md

LLM failover automatically retries a failed request on a backup so a single model or backend outage doesn't surface to your users as an error.

Models and their backends have outages, rate limits, and slow spells. Without failover, any of these becomes a user-facing error and your application has to implement its own retry-and-fallback logic. Failover moves that resilience into the gateway.

A robust failover chain has several layers: retry the same model, then try an equivalent model on a different backend, then fall back to a quality-equivalent model. Done well, the user sees a slightly slower response instead of an error, and you write none of the retry logic.

## How Kyma does it

Kyma uses a multi-layer never-die chain (retry, equivalent-model swap, then a parallel race across quality-tier peers), so user-facing errors stay near zero. When a fallback serves, response headers tell you which model and which layer handled it.

- [Live status](https://kymaapi.com/status)
- [What is an LLM gateway?](https://kymaapi.com/glossary/llm-gateway)
- [Compare](https://kymaapi.com/compare)

Try it with $0.50 free credit on the free tier. No card required. · [Get API key →](https://kymaapi.com/signup)

More terms

- [LLM gateway](https://kymaapi.com/glossary/llm-gateway)
- [LLM router](https://kymaapi.com/glossary/llm-router)
- [Prompt caching](https://kymaapi.com/glossary/prompt-caching)
- [Model aliases](https://kymaapi.com/glossary/model-aliases)
- [Tool calling](https://kymaapi.com/glossary/tool-calling)
- [Structured outputs](https://kymaapi.com/glossary/structured-outputs)
- [Pay-per-token pricing](https://kymaapi.com/glossary/pay-per-token)