Cogito v1 Preview 8B

Good fit for

People comparing direct responses with opt-in reasoning in the same locally run model.

Why consider it

The exact provider card documents normal mode, a system instruction for extended thinking and structured tool-calling examples.

Main limitations

The upstream 8B card identifies Llama 3.1 licensing, while the runtime embeds Llama 3.2 license text. Inspect the actual downloaded artifact and upstream terms before deployment. Tool calls still require an application to execute them; no code execution or web-search feature is implied. Historical benchmark superiority claims are omitted.

Cost and access

No hosted subscription or token fee for local inference within the license’s permitted uses. Hardware, electricity and setup are separate. Provider weights are documented under the Llama 3.1 Community License; the runtime license-label discrepancy is disclosed rather than treated as legal clearance.

Getting started

Install a current compatible Ollama version, review the exact model license, then run: ollama run cogito:8b. API means the local runtime endpoint, not a hosted service included with the weights.