NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-10-05-01M46B33D5GPET0075301BNS08
Run Details
Model:
GPT-6.1 Sol (Reasoning, Medium)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
1m 13s
Cost:
$0.022822
Provider:
Azure
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
80.81%
Dialogue to Total Word Ratio
(details)
90.4060%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 70% words of dialogue between two thieves named Molly and Bill who are planning to break into a children's party to steal all the presents. Only include the prose of the scene without any markdown or commentary.
Model response
Messages
Meta
AI
Molly spread the birthday invitation across the seat of their stolen shopping trolley. Rain had blurred the clown into something with outstanding warrants. Bill studied it through binoculars, although he was standing six inches away. Beside them, a suitcase full of counterfeit balloon animals squeaked whenever either thief breathed confidently. “We go in as entertainers,” Molly said. “You distract the children. I collect the presents. We leave before cake. Clean, simple, practically respectable. Have you brought the clown costume, or should I expect another incident?” “I brought it,” Bill said. “But the trousers have a security tag. Every time I pass a supermarket, my backside confesses. Also, I can’t juggle. I can disappoint three oranges simultaneously, which looks fairly similar.” “Fine. You’re a magician. Magicians are expected to lose things. Start with your dignity. Then produce a rabbit, vanish a handkerchief, and keep everyone looking away from the gift table. What tricks do you know?” “I can find somebody’s card eventually,” Bill said. “Provided they tell me which card it is, and nobody has shuffled. Once I made a parking ticket disappear. The council described that trick as destroying evidence.” Molly took a pencil from behind her ear and drew a floor plan on the invitation. She marked the gift table with an X, the kitchen with a crown, and the birthday child with a small, deeply unflattering tornado. Bill added an emergency exit shaped suspiciously like a custard tart. “That exit is a pastry,” Molly said. “Focus. There will be twelve children, two parents, and a grandmother. The grandmother is our main concern. Grandmothers can identify professional dishonesty at forty paces, even through walls.” “Mine caught me stealing biscuits over the telephone. She hadn’t even called our house. Wrong number. Terrifying woman. Should we take flowers? Criminals with flowers look less criminal. Unless the flowers are also stolen.” “No flowers. We need bags. Big ones. We take every present, including anything that rattles. Especially anything that rattles. Rich children get expensive construction sets. Poor thieves deserve a second chance at structural engineering, Bill.” “What about the cards? Last time you stole a birthday envelope, it contained fifty pounds and a strongly worded reminder to moisturise. I’m still using the cream. My elbows have never felt so morally complicated.” A whistle shrilled beyond the garden fence. Neither thief moved. Through a gap, they could see a dozen children wearing pirate hats. Each held a foam sword. Beside the presents stood a grandmother with folded arms and the expression of a woman who had personally invented consequences. Bill swallowed hard. “New plan,” Molly whispered. “You approach the grandmother. Compliment her cardigan. Ask about the war. Any war. While she corrects your assumptions, I slip behind the pirates and gather the loot. Try not to surrender.” “Molly, that’s my grandmother. This is my nephew’s party. I wondered why the address sounded familiar. Good news: she likes you. Bad news: she knows we’re coming. She asked me to bring bin bags yesterday.”
created_at
"2026-10-05T15:32:05.167Z"
10/5/2026, 3:32:05 PM
model
"openai/gpt-6.1-sol-20260929"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
14096
moderation_latency
(null)
generation_time
72389
tokens_prompt
75
tokens_completion
1242
native_tokens_prompt
66
native_tokens_completion
2269
native_tokens_completion_images
0
native_tokens_reasoning
1619
native_tokens_cached
0
num_media_prompt
(null)
num_input_audio_prompt
(null)
num_media_completion
0
num_search_results
(null)
num_fetches
(null)
web_search_engine
(null)
origin
"https://ncbench.com/"
is_byok
false
finish_reason
"stop"
native_finish_reason
"completed"
service_tier
"default"
usage
0.022822
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1791214325-KGIf1Kpl7NAKOkz26fIT"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1791214325-Xapb75J2dTK2g4k1P6vl"
upstream_id
"1413d8b4-c3cf-4a76-a14e-2b7331485e2e"
provider_responses
0
endpoint_id
"dfb9d5c3-63f9-4263-8b78-4fa97dcb3d5e"
id
"resp_09737c2166bb6911016ac3c2f549d487d1b32ba78adaeee04d"
is_byok
false
latency
420
model_permaslug
"openai/gpt-6.1-sol-20260929"
provider_name
"OpenAI"
status
502
1
endpoint_id
"32e6ec1d-6c62-4d75-bd38-1b3899c4db8b"
id
"1413d8b4-c3cf-4a76-a14e-2b7331485e2e"
is_byok
false
latency
550
model_permaslug
"openai/gpt-6.1-sol-20260929"
provider_name
"Azure"
status
200
total_cost
0.022822
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Azure"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
100.00%
Matches word count
n/a
neededClean
false
words
499
80.81%
Dialogue to Total Word Ratio
Ratio: 67.94%, Deviation: 2.06%
neededClean
false
wordsTotal
499
wordsDialogue
339
90.4060%