September 29, 2026
Agent traces can now be evaluated on OpenRouter: Span-01 and Lite available
On Sep 28, OpenRouter made Span-01 and Span-01 Lite available for evaluating agent traces. The models return the probability of a specified attribute for a span: for example, whether a user is frustrated or a tool call is safe.
OpenRouter
@openrouter
Span-01 and Span-01 Lite from @RespanAI are now available in OpenRouter. These models evaluate agent traces. Send them a span and the attributes that matter to you, and get the probability of each one: for example, “is the user frustrated?” or “is this tool call safe to run?”
Respan @RespanAI
Introducing Span-01, the first hyper-parallel reasoning classifier built for unseen challenges (RLAIF). 2x cheaper, 18% better than Jev. 700x cheaper, 4% better than GPT-6 Luna. Frontier reasoning for every behavior, at classifier speed. • Span-01: #1 on Behavior Benchmark • Span-01 Lite: Better than Jev and completely free!
· 17.4K views
On Sep 28, OpenRouter made Span-01 and Span-01 Lite available for evaluating agent traces. The models return the probability of a specified attribute for a span: for example, whether a user is frustrated or a tool call is safe.
RespanAI calls Span-01 Lite completely free.
