Working paper / 2026
Conversational AI LLM Bench
Abstract
We define conversational AI LLMs as the language models at the center of cascaded voice pipelines, between speech recognition and speech synthesis, that return a first token within one second: a median time to first token of 1,000 ms or less. The Conversational AI LLM Bench (CALLM) benchmarks this category. Each model operates every agent of a multi-agent front desk in healthcare, legal, and customer support, with production-length prompts, executable tools, and per-call database state. A simulated caller pursues a locked task over a simulated phone line, so the model receives only what a speech recognizer would have transcribed, errors included. A deterministic verifier scores tool calls, handoffs, and final database state; pass@1 and pass5 measure single-attempt capability and five-attempt reliability.
Clone the repository, point a model at an industry pack, and reproduce the evaluation environment.
Healthcare / C3-H2
What this task measures
The caller is on the line with Straus Dermatology for one locked request: Out-of-pocket cost without a member ID. They keep repeating that ask through a waitlist offer, a callback, or a different errand. We included it as a hard coverage and benefits case because carrier answers have to stay tied to one office and one plan, naming that location and that carrier. The agent succeeds only if it stays on that request, reaches the outcome by following the coverage desk then scheduling path, and leaves the record exactly as this case requires: the right row written or cleared, and nothing extra.
One model at a time, with a representative pass and fail conversation when both exist. GPT-4.1 is the default view.
Conversation cost $0.062
0:00 / 0:00 · scrub · trace sample · click a span to play, drag to scrub, scroll to pan
Audio
Speech Audio
Agent Audio
Handoffs
Tool Calls
Errors
Transcript
How a conversation passes
Combined pass is the logical and of three deterministic gates. Tools, handoff path, and final database state must all pass on the same conversation. One miss fails the call.
Combined
Fail
This sample fails because tools and handoff path and final database state did not match. Combined cannot pass unless every gate does.
Expected versus actual tool calls on this conversation.
Expected
Actual
Parameters
None recorded.
None recorded.
Output
None recorded.
None recorded.
Parameters
- carrier: cigna
- location_id: loc_park_ave
Output
None recorded.
None recorded.
Parameters
None recorded.
None recorded.
Output
+ role: scheduling
+ success: true
Parameters
is_new_patient: true
is_new_patient: true
+ urgency: routine
visit_class: medical
visit_class: medical
Output
+ data: {"appointment_type_code":"NP_MED","visit_class":"medical","required_credential":"MD_OR_PA","duration_min":30,"urgency":"routine","constraints":[],"is_new_patient":true}+ error_code: empty
+ ok: true
+ patient_safe_message: empty
Parameters
location_id: loc_park_ave
location_id: loc_park_ave
Output
+ data: {"locations":[{"id":"loc_park_ave","name":"Park Avenue","zip":"10016","offers_cosmetic":true,"address":"386 Park Avenue South, New York, NY 10016","floor":"4th floor","suite":"Suite 410","hours":"Mon-Fri 8am-5pm, Sat 9am-1pm","services":"medical, surgical, cosmetic, allergy","transit":"One block from the 6 train at 33rd Street","parking":"Garage next door on East 27th"},{"id":"loc_brooklyn_heights","name":"Brooklyn Heights","zip":"11201","offers_cosmetic":true,"address":"142 Montague Street, Brooklyn, NY 11201","floor":"2nd floor","suite":"Suite 2B","hours":"Mon-Fri 8am-5pm","services":"medical, surgical, cosmetic, allergy","transit":"Two blocks from Borough Hall","parking":"Street parking; garage on Clinton Street"},{"id":"loc_windermere","name":"Windermere","zip":"34786","offers_cosmetic":false,"address":"7600 Conroy Windermere Road, Windermere, FL 34786","floor":"Ground floor","suite":"Suite 100","hours":"Mon-Fri 8am-5pm","services":"medical, surgical, allergy","transit":"No transit nearby","parking":"Free surface lot"}]}+ error_code: empty
+ ok: true
+ patient_safe_message: empty
Parameters
location_ids: ["loc_park_ave"]
location_ids: ["loc_park_ave"]
Output
+ data: {"slots":[{"slot_id":"slot_loc_park_ave_1","location_id":"loc_park_ave","location":"Park Avenue","provider_id":"prov_chen","provider":"Dr. Amy Chen, MD, FAAD","start":"2026-08-24T09:00","end":"2026-08-24T09:30"},{"slot_id":"slot_loc_park_ave_2","location_id":"loc_park_ave","location":"Park Avenue","provider_id":"prov_chen","provider":"Dr. Amy Chen, MD, FAAD","start":"2026-08-25T11:30","end":"2026-08-25T12:00"},{"slot_id":"slot_loc_park_ave_3","location_id":"loc_park_ave","location":"Park Avenue","provider_id":"prov_chen","provider":"Dr. Amy Chen, MD, FAAD","start":"2026-08-26T14:00","end":"2026-08-26T14:30"}],"count":3}+ error_code: empty
+ ok: true
+ patient_safe_message: empty
Parameters
appointment_type_code: NP_MED
appointment_type_code: NP_MED
end: 2026-08-24T09:30
end: 2026-08-24T09:30
location_id: loc_park_ave
location_id: loc_park_ave
provider_id: prov_chen
provider_id: prov_chen
slot_id: slot_loc_park_ave_1
slot_id: slot_loc_park_ave_1
start: 2026-08-24T09:00
start: 2026-08-24T09:00
Output
+ data: {"appointment":{"id":28,"patient_id":null,"location_id":"loc_park_ave","provider_id":"prov_chen","appointment_type_code":"NP_MED","start":"2026-08-24T09:00","end":"2026-08-24T09:30","description":"new patient visit","status":"booked","created_at":"2026-11-11 01:08:53"},"status":"booked"}+ error_code: empty
+ ok: true
+ patient_safe_message: empty
Parameters
- appointment_type_code: NP_MED
- earliest: 2026-08-24
- latest: 2026-09-30
- location_ids: ["loc_park_ave"]
Output
None recorded.
None recorded.
Parameters
+ reason: caller_done
Output
+ success: true
Open a row for parameters and outputs. Required tools must appear. Extra industry tools are allowed. Missing: transfer_to_coverage, check_plan_accepted, join_waitlist. Extra: end_call.
Full industry architecture, same graph as the industry page. The left diagram marks the expected specialist path in green. The right diagram marks the path this conversation actually took in green through the shared prefix, then red from the first wrong hop.
Expected path
transfer_to_coverage → transfer_to_scheduling
Actual path
transfer_to_scheduling
Rows that differ between the expected hangup and the actual hangup.
missing · waitlist · 1
Expected
Actual
- appointment_type_code: NP_MED
- earliest: 2026-08-24
- id: 1
- latest: 2026-09-30
- location_ids: ["loc_park_ave"]
- patient_id: empty
The locked caller script. Copy it as written.
You are a new patient with no chart at Straus. You do not have your insurance card with you, and you will not get it during this call. In your first real answer, say you have Cigna without your card and ask whether they take Cigna at Park Avenue and what a visit will cost out of pocket there. If they cannot give a number, ask a second time for any rough ballpark, then say you still want to be booked for a routine rash at Park Avenue. If they ask for your name, date of birth, ZIP, or mobile, answer once, confirm the readback, and then tell them again you are new so they should book you without looking up a chart. If they ask what insurance you have, say Cigna; you do not have a member ID or group number and will not make one up. Take the first appointment time they offer. After that time is on the calendar, ask to be waitlisted at Park Avenue from August twenty-fourth through September thirtieth in case a better opening comes up, those calendar dates only. Never call that range earlier dates. Confirm that first time out loud, then thank them and end the call. Do not ask to add insurance, for a callback, to speak to a person on this call, about a bill you already owe, a payment, the patient portal, a text, financing, or the street address, floor, or suite. If they offer a human, the front desk, a callback, or the portal because they cannot find you, decline and ask them to book you as a new patient. A listed opening is not a booking. Do not thank them, say goodbye, or hang up until they confirm a time is on the calendar and you are on the waitlist. If they only list openings, say book the first one. Do not ask them to text you an appointment confirmation. When they tell you something and then ask a question, answer the question in the same reply. Once they have said the time is booked and you are on the waitlist, confirm both once, thank them, and end the call; do not ask them to confirm again.