τ-bench

Source-qualified evaluations, report statistics and safety observations.

gpt-realtime-1.0

MetricValueUnit
τ-bench · airline · pass^10.36percent
τ-bench · retail · pass^10.35964912280701755percent
τ-bench · telecom · pass^10.19298245614035087percent

gpt-realtime-1.5

MetricValueUnit
τ-bench · airline · pass^10.4percent
τ-bench · retail · pass^10.447percent
τ-bench · telecom · pass^10.21100000000000002percent

gpt-realtime-2

MetricValueUnit
τ-bench · banking_knowledge · pass^10.10309278350515463percent

gpt-realtime-2

MetricValueUnit
τ-bench · airline · pass^10.56percent
τ-bench · retail · pass^10.39473684210526316percent
τ-bench · telecom · pass^10.20175438596491227percent

gpt-realtime-2

MetricValueUnit
τ-bench · airline · pass^10.66percent
τ-bench · retail · pass^10.5percent
τ-bench · telecom · pass^10.37719298245614036percent

gpt-realtime-2

MetricValueUnit
τ-bench · airline · pass^10.58percent
τ-bench · retail · pass^10.47368421052631576percent
τ-bench · telecom · pass^10.21929824561403508percent

Grok Voice 3.0

MetricValueUnit
τ-bench · airline · pass^10.8percent
τ-bench · retail · pass^10.7631578947368421percent
τ-bench · telecom · pass^10.8245614035087719percent

grok-voice-think-fast-1.0 + tool-mentor

MetricValueUnit
τ-bench · airline · pass^10.7percent
τ-bench · retail · pass^10.7631578947368421percent
τ-bench · telecom · pass^10.7807017543859649percent

grok-voice-think-fast-1.0

MetricValueUnit
τ-bench · airline · pass^10.66percent
τ-bench · retail · pass^10.6228070175438597percent
τ-bench · telecom · pass^10.7368421052631579percent

grok-voice-think-fast-2.0

MetricValueUnit
τ-bench · airline · pass^10.56percent
τ-bench · retail · pass^10.5964912280701754percent
τ-bench · telecom · pass^10.7192982456140351percent

Cascaded baseline

MetricValueUnit
τ-bench · airline · pass^10.48percent
τ-bench · retail · pass^10.2894736842105263percent
τ-bench · telecom · pass^10.16666666666666663percent

phonic-1

MetricValueUnit
τ-bench · airline · pass^10.8percent
τ-bench · retail · pass^10.7631578947368421percent
τ-bench · telecom · pass^10.9473684210526315percent

phonic-1

MetricValueUnit
τ-bench · airline · pass^10.9percent
τ-bench · retail · pass^10.7982456140350878percent
τ-bench · telecom · pass^10.9824561403508771percent

Pine Voice 1.0

MetricValueUnit
τ-bench · airline · pass^10.92percent
τ-bench · retail · pass^10.7876106194690266percent
τ-bench · telecom · pass^10.9122807017543859percent

Pine Voice Preview

MetricValueUnit
τ-bench · airline · pass^10.8percent
τ-bench · retail · pass^10.8508771929824561percent
τ-bench · telecom · pass^10.7543859649122807percent

Pine Voice Preview

MetricValueUnit
τ-bench · airline · pass^10.7percent
τ-bench · retail · pass^10.7017543859649122percent
τ-bench · telecom · pass^10.8596491228070176percent

qwen3.5-omni-plus-realtime

MetricValueUnit
τ-bench · airline · pass^10.54percent
τ-bench · retail · pass^10.4649122807017544percent
τ-bench · telecom · pass^10.6052631578947368percent

Salesforce Launa v1

MetricValueUnit
τ-bench · airline · pass^10.66percent
τ-bench · retail · pass^10.7192982456140351percent
τ-bench · telecom · pass^10.8070175438596491percent

Models

Loading
EVIDENCE / COMPARISON

Benchmarks

Explore every benchmark

Loading

These charts contain model-linked results. The charts above also include unlinked source profiles. Compare scores within each evaluation. Model quality, usage, transparency and system performance measure different things.

Install TensorTrace