RespanAI announced that its open-weight decision model Span-02 is now available on OpenRouter, supporting text and image inputs and returning a choice, yes/no/unknown, or a score with probabilities. It is priced at $0.01 per million input tokens with free output, a 131K context window, and a Rust inference engine that is 7x faster than vLLM.
RespanAI's decision models Span-01 and Span-01 Lite are now available on OpenRouter for agent-trace behavior analysis, taking a trace span and targeted behaviors as input to return probability scores. The company says Span-01 is a hyper-parallel reasoning classifier for unseen challenges, with lower cost and higher performance than Jev and GPT-6 Luna, ranking first on the Behavior Benchmark, while Span-01 Lite beats Jev and is free.