1016
ELOELO Rating
8th of 12
24.3%
Pass³Reliability (Pass³)
9th of 12
2.10s
LatencyLatency P50
6th of 12
Higher wins more head-to-head calls (1000 = average); further left replies faster
Each model took the same recorded human callers on live calls. Pick a measure to see where Gemini 3.8 Live Extended Thinking ranks.
Rating from head-to-head calls: on every call each pair of models is compared, the better call wins. 1000 is the average model. Higher is better.
Pick a skill family. The shape is this model on every skill in it; the dashed line is the median of the other models.
Share of calls where the skill was needed and handled, strongest first. Hover a point on the chart for both values.
Banking and healthcare callers. Each shows what the caller needed, what made the call hard, and how the model handled it.
Finished 4 of 6 steps
Handled well: Overlapping-Speech Recovery, Human Handoff, Redundant Re-prompt Avoidance
Struggled with: Emotion Recognition, Multi-Speaker Handling, Tool-Latency Handling
Listen to the callFinished 2 of 3 steps
Handled well: TTS Text Normalization, Language Switching
Listen to the callFinished 6 of 7 steps
Handled well: Multi-Speaker Handling, Overlapping-Speech Recovery, Redundant Re-prompt Avoidance
Struggled with: Degraded-Channel Robustness
Listen to the callFinished 6 of 7 steps
Handled well: Degraded-Channel Robustness, Backchannel vs Barge-in Detection, Consent Verification
Listen to the callFinished 3 of 6 steps
Handled well: Mid-Utterance Pause Tolerance, Consent Verification, Redundant Re-prompt Avoidance
Struggled with: Low-Confidence Confirmation, Prosodic Cue Understanding
Listen to the callFinished the task
Handled well: Waiting Through Hold, Multi-Speaker Handling, Backchannel vs Barge-in Detection
Listen to the callFinished 0 of 3 steps
Handled well: Tool-Latency Handling, Mid-Utterance Pause Tolerance, Redundant Re-prompt Avoidance
Struggled with: Premature-Response Avoidance, Consent Verification, Numeric Read-back
Listen to the callFinished 1 of 2 steps
Handled well: Consent Verification, Cancellation Handling, Redundant Re-prompt Avoidance
Listen to the callGemini 3.8 Live Extended Thinking starts replying in 2.10s on a typical turn and 2.86s on the slowest turn in twenty, 6th fastest of 12 at the median.
210 live calls: 70 scenarios, each run three times. The median call ran 4:53.
70 scenarios, the same for every model.
Gemini 3.8 Live Extended Thinking has an ELO rating of 1016, 8th of 12 ranked models. It passed 17 of 70 callers on all three runs, a Pass³ of 24.3%.
Gemini 3.8 Live Extended Thinking has an ELO rating of 1016, 8th of 12 ranked models. It passed 17 of 70 callers on all three runs, a Pass³ of 24.3%.
Gemini 3.8 Live Extended Thinking starts replying 2.10s after the caller finishes at P50, 2.60s at P90 and 2.86s at P95: 6th fastest of 12 at P50, measured on the call audio.
Gemini 3.8 Live Extended Thinking falls furthest behind on Corrections and Changes: 40%, against a median of 61%.
Gemini 3.8 Live Extended Thinking has a task completion rate of 22.9%, 7th of 12. A call counts only when every step the caller needed was done.
There is no verified per-minute rate for Gemini 3.8 Live Extended Thinking, so it is left out of the cost ranking.
Cekura ran Gemini 3.8 Live Extended Thinking as the whole agent with the same prompt, tools and phone connection as every other model. It took the same 70 recorded real human callers on live calls, three times each (210 calls), and each call was graded on 25 skills across Hearing, Turn Taking, Task Handling and Speaking.
Cekura runs these simulated calls on your agent and scores every one.
