NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-04-24-01KQ0PXV610Z5X4BAFQ0RFVDRK
Run Details
Model:
GPT-5.5 (Reasoning, Low)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
28.8s
Cost:
$0.034200
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.03%
Dialogue to Total Word Ratio
(details)
0.0129%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 50% 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 crayon-drawn map across the café table and pinned down one curling corner with her teacup. The map showed a church hall, a bouncy castle, six balloons, and a large rectangle labeled PRESENT MOUNTAIN in purple wax. Bill leaned over it, squinting. “Why is the exit labeled ‘dragon’?” “That’s not an exit. That’s the entertainer.” “The entertainer is a dragon?” “No, Bill, he’s a man in a dragon costume. Try to keep up.” Bill nodded solemnly and stirred three sugars into his coffee. “Right. Man-dragon. Classic security.” Around them, lunchtime customers ignored the two most suspicious people in town. Molly wore dark glasses indoors. Bill had a party hat sticking out of his coat pocket. “The plan is simple,” Molly said. “We enter at fourteen hundred hours disguised as responsible adults.” Bill frowned. “Can we pull that off?” “Not convincingly, but children are easily distracted.” “I was a child once,” Bill said. “I noticed everything.” “You ate paste until you were nine.” “Exactly. I knew where the paste was.” Molly tapped the PRESENT MOUNTAIN. “The birthday boy is called Toby. Seven years old. Loves dinosaurs, magic tricks, and shouting. There will be thirty guests, four parents, one grandmother with eyes like a hawk, and the dragon.” Bill shivered. “The grandmother worries me.” “She should. Her name is Doris. She once caught me taking two biscuits from a funeral buffet.” “You stole from a funeral?” “I was grieving in advance.” Bill pulled a small notebook from his pocket. “So, how do we get near the presents?” “We volunteer to help.” “With what?” “Anything. Plates. Juice. Emotional support. Children always need mopping.” Bill wrote that down. “Mopping children.” “Do not write that down.” He crossed it out. “What about the presents? Big sacks? Van outside? Smoke bomb?” “No smoke bomb.” “Glitter bomb?” “No bombs.” “Tiny trumpet?” Molly stared at him. “For distraction,” Bill explained. “Everyone hears a tiny trumpet, everyone turns.” “No tiny trumpet.” Bill looked wounded and slid the trumpet deeper into his coat. Molly lowered her voice. “We wait until the cake comes out. Everyone sings. Eyes on candles. That is our moment.” “To steal every present from a child.” “Yes.” Bill chewed his pencil. “When you say it plainly, it sounds mean.” “It’s a heist, Bill. Heists are rarely praised by school inspectors.” “But they’re children’s presents. There might be socks.” “There might be electronics.” “There might be a handmade card from Nana.” Molly hesitated. “We don’t steal cards.” “How will we know?” “We shake things.” Bill brightened. “I’m excellent at shaking things. Once identified a jigsaw puzzle through wrapping paper.” “That was because it said ‘Jigsaw Puzzle’ on the tag.” “Deduction has many faces.” A waiter passed by, glanced at the map, and kept walking faster. Molly folded the map. “Remember the roles. I’m Auntie Margaret.” “You don’t look like an Auntie Margaret.” “I have pearls.” “They’re plastic.” “So is Auntie Margaret’s smile.” Bill lifted his hand. “Who am I?” “You are Clive, the magician’s assistant.” “I thought there was a dragon.” “There is. You assist him.” “How?” “By staying out of the way.” Bill considered this. “I can do that badly.” “Yes, I know.” Outside the café window, a bus went past advertising Happy Hippo Party Hire. A painted hippo grinned down at them with unsettling confidence. Bill sighed. “Molly?” “What?” “Did you have birthday parties as a child?” She stiffened. “Of course.” “With presents?” “Sometimes.” “With cake?” “Cake-adjacent items.” “That sounds like bread.” “It was festive bread.” Bill placed the pencil down. “Maybe instead of stealing the presents, we steal the leftover sausage rolls.” Molly blinked. “That is not a heist.” “It could be. Operation Sausage Roll.” “No.” “And we could add a present.” “Add?” He pulled a wrapped parcel from his bag, crookedly taped and covered in penguin paper. “I bought a dinosaur. For camouflage.” Molly looked at the parcel, then at Bill. “You bought a present for the child we’re robbing?” “It felt rude to arrive empty-handed.” For a long moment, Molly said nothing. Then she unfolded the map again and crossed out PRESENT MOUNTAIN. In its place she wrote SAUSAGE ROLLS in stern purple letters. Bill smiled. “So we’re good people?” “Absolutely not,” Molly said, standing. “We are criminals with boundaries.” “And a tiny trumpet?” “One note,” she warned. Bill sprang up, delighted, and gave the smallest toot the café had ever heard. Every customer turned. Molly sighed. “Fine. It works.”
created_at
"2026-04-24T21:39:25.512Z"
4/24/2026, 9:39:25 PM
model
"openai/gpt-5.5-20260423"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3413
moderation_latency
(null)
generation_time
28748
tokens_prompt
77
tokens_completion
1430
native_tokens_prompt
66
native_tokens_completion
1129
native_tokens_completion_images
0
native_tokens_reasoning
77
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"
usage
0.0342
router
(null)
provider_responses
0
endpoint_id
"58e5b336-423e-430b-a2ab-8bc353f0c51b"
id
"resp_071066aa33c005da0169ebe30d9a1c8195b6427f229ecd2eba"
is_byok
false
latency
270
model_permaslug
"openai/gpt-5.5-20260423"
provider_name
"OpenAI"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.1; linux; x64))"
http_referer
(null)
request_id
"req-1777066765-0ohrQ9poANwjVFY9KRSb"
session_id
(null)
api_type
"completions"
id
"gen-1777066765-rBggIShA4aarKlhaMROG"
upstream_id
"resp_071066aa33c005da0169ebe30d9a1c8195b6427f229ecd2eba"
total_cost
0.0342
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
740
0.03%
Dialogue to Total Word Ratio
Ratio: 62.85%, Deviation: 12.85%
neededClean
false
wordsTotal
743
wordsDialogue
467
0.0129%