Datadog builds specialist AI model Scout for SRE investigations

Datadog released Scout, a model trained to investigate production incidents inside its Bits Investigation AI SRE product, analyzing alerts and telemetry to suggest root causes. On Datadog's AlertEval benchmark, Scout medium matched GPT-6.1 Sol high and Claude Opus 5.5 high at 3–8x lower inference cost. It is in closed preview with Datadog engineers and select customers.

• Scout is based on GLM-5.3 Flash and post-trained with reinforcement learning
• On SREGym Scout resolved 71, 79 and 80 of 90 faults at medium, high and xhigh
• GPT-6.1 Sol xhigh completed 82 of 90 cases on the same set
• No GA date announced; Scout diagnoses and proposes fixes but does not apply them