Agentic Inference Benchmarks
InferenceX agentic and fixed-sequence AI inference benchmarks compare throughput (tok/s/chip), interactivity (tok/s/user), time to first token (TTFT) and cost across chips, serving frameworks and models. Agentic results come from AgentX, which replays long-context, multi-turn coding-agent traces with subagents. Every data point comes from a public GitHub Actions workflow, so it is reproducible and auditable.
Inference Performance
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.