NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M37HGGYVB7DVS842NEDR6SHC
Run Details
Model:
GPT-6 Sol
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
16.5s
Cost:
$0.007292
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
0.03%
Matches word count
(details)
79.10%
Dialogue to Total Word Ratio
(details)
39.5628%
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 a crayon drawing across the café table. It showed a purple castle, seventeen stick figures, and a dog wearing a crown. “Is that the target?” Bill asked. “That’s the invitation.” “Why is the dog king?” “Focus. Saturday, two o’clock. The twins turn seven. Thirty guests, two entertainers, and a mountain of presents.” Bill leaned closer. Someone had written YOU’RE INVITED! in glitter glue. A fleck caught his sleeve and refused to leave. “We’re stealing from children,” he said. “We’re stealing unopened presents. There’s a distinction.” “What distinction?” Molly studied the drawing. “One sounds worse.” Outside, rain polished the street. Inside, the café’s espresso machine shrieked like a witness identifying suspects. Molly rotated the invitation toward Bill and tapped the castle with a chipped fingernail. “The party’s at Bounce Kingdom. Presents go on a table beside the inflatable dragon. We enter as entertainers, create a distraction, and leave with the goods.” “What sort of entertainers?” “Balloon artists.” Bill’s expression tightened. “I can make a snake.” “That’s enough.” “It doesn’t look like a snake.” “Then call it a worm.” “It doesn’t look like a worm either.” Molly folded the invitation before the glitter spread. She had acquired it from her sister’s fridge, along with a suspicion that her sister would recognize her immediately. “You’ve been there before?” Bill asked. “Once. The dragon has two exits. The rear one opens by the car park.” “Good. We’ll have an escape route.” “It’s an emergency exit. We’re not using it unless there’s an emergency.” Bill looked at the dog’s crown. “How much is all this worth?” “No idea. But I saw the gift registry. Remote-control trucks, building sets, a karaoke machine.” “Couldn’t we steal the registry instead?” “It’s a list, Bill.” “We could use it to choose a less crowded shop.” For a moment Molly considered that. Then she slid him a packet of modelling balloons and watched him pull out a yellow one. He stretched it twice, blew until his cheeks reddened, and produced a noise like a duck stepping on another duck. “It has a hole,” he announced. “Try another.” Bill selected blue. This time the balloon swelled, snapped free of his fingers, and ricocheted under three tables. A woman retrieved it from her handbag while he apologised. Molly lowered her voice. “We also need matching shirts.” “I have a shirt.” “It says SECURITY on the back.” “People trust security.” “Not security carrying sacks of birthday presents.” Bill gathered the limp balloon from the floor. On its return journey, it had picked up a crumb and a strand of hair. “Why did your sister invite you?” he asked. “Because I’m the twins’ aunt.” He stared at her. “They’ll know your face.” “Yes. I’m bringing the cake.” “Then who’s stealing the presents?” Molly paused. Her sister had asked for help loading the car after the party; that was why the invitation was in her bag. She pressed a thumb into the glittery crease. “Neither of us,” she said. “You’re coming to help carry them.” Bill considered the balloon. “Can I still make a snake?” “For the twins.” He twisted it carefully. It became a convincing sword, which Molly decided not to mention.
created_at
"2026-09-23T16:27:49.093Z"
9/23/2026, 4:27:49 PM
model
"openai/gpt-6-sol-20260922"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3120
moderation_latency
(null)
generation_time
16417
tokens_prompt
75
tokens_completion
917
native_tokens_prompt
66
native_tokens_completion
716
native_tokens_completion_images
0
native_tokens_reasoning
0
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
0.007292
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.20.0; linux; x64))"
http_referer
(null)
request_id
"req-1790180869-B2q6sW4pAKg0VnuaPiFu"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790180869-4iagsJ1H4ekOsfiuQsPw"
upstream_id
"resp_0eafb49a8065f86b016ab3fe054df887d18fab8f43bdaa2f81"
provider_responses
0
endpoint_id
"1cfc7d9d-4404-4b8e-9ee4-58ae45c9dcd4"
id
"resp_0eafb49a8065f86b016ab3fe054df887d18fab8f43bdaa2f81"
is_byok
false
latency
1611
model_permaslug
"openai/gpt-6-sol-20260922"
provider_name
"OpenAI"
status
200
total_cost
0.007292
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
0.03%
Matches word count
n/a
neededClean
false
words
530
79.10%
Dialogue to Total Word Ratio
Ratio: 47.83%, Deviation: 2.17%
neededClean
false
wordsTotal
531
wordsDialogue
254
39.5628%