Which LLM does Creopus use?

Creopus runs a provider-fallback chain, configurable per tool:

Tool	Default order
Most tools (SWOT, DFMEA, DVP, Pugh, RCCA, ReqGen, Flowchart)	Cerebras → Gemini 2.5 Flash → Claude
Multimodal / large context (BOM Consolidation, PCBGen, Design Doc)	Gemini 2.5 Flash → Claude → Cerebras

If the primary provider is rate-limited, errors out, or the prompt exceeds its context window, the platform automatically retries on the next provider with no user-visible delay beyond a few hundred ms.

Enterprise customers can pin a single provider, supply their own API keys (BYOK), or deploy a self-hosted open-weight model.