NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-09-01KX408G8M5V5FM0R98E48EGTD
Run Details
Model:
GPT-5.6 Terra
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
12.7s
Cost:
$0.014415
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
0.0000%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 30% 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 crouched behind the hedge with the patience of a saint and the knees of a woman who had made several poor career decisions. Beyond the garden fence, fairy lights blinked around a pink-and-blue gazebo. Children screamed with the reckless joy of people too small to understand tax law. Bill peered through a gap in the leaves. “That’s a lot of presents.” “That,” Molly whispered, “is why we’re here.” “There must be thirty boxes.” “Thirty-two. I counted the gift bags. Two of them are shaped like unicorns, which means they’re either expensive toys or a warning from God.” Bill adjusted the black wool cap pulled over his ears. “I’m still uneasy about this.” “You were uneasy stealing the mayor’s watch.” “The mayor was insured.” “So are these parents. Probably. They look like the sort who insure patio furniture.” A woman in a tiara clapped her hands near the gazebo. “Cake in ten minutes!” The children cheered as if she had announced the discovery of a new planet made entirely of sugar. Bill swallowed. “Molly, these are children.” “They’re wealthy children.” “That doesn’t make it better.” “It makes it organized. Look at the wrapping paper. Coordinated ribbons. No child picked those colors. This is a neighborhood where people say ‘curated’ about sandwiches.” Bill watched a boy in a pirate costume whack a cardboard dragon with a foam sword. “What exactly are we stealing, then?” “Presents.” “All of them?” “All of them.” “What if there’s a bicycle?” “We take the bicycle last.” “What if there’s a pony?” “Then we abandon the operation and become horse thieves, which is a separate business plan.” Molly unfolded a hand-drawn map. It had been sketched on the back of a takeaway menu and smelled faintly of curry. “The presents are on that table by the gazebo. The adults are near the barbecue, drinking pink wine out of jars. The children are occupied by a magician named Sparklebeard.” Bill squinted. “Sparklebeard?” “He has a beard with glitter in it. He’s doing balloon animals.” “That man is braver than us.” “He’s charging three hundred pounds for an hour.” “Then he’s braver and smarter.” Molly pointed to the side gate. “We enter there. You distract the magician.” “I distract a magician?” “You’re good at being confused.” “I’m not confused.” “You asked whether a pony would be in a gift bag.” “That was a practical question.” From the garden came a burst of applause. Sparklebeard produced a bouquet of flowers from a child’s shoe. The child stared at it, deeply offended. Bill rubbed his hands together. “What’s my distraction?” “Walk up, announce that you’re the birthday girl’s uncle, and ask if anyone has seen your ferret.” “I don’t have a ferret.” “Exactly. It raises questions without providing answers.” “And while everyone looks for an imaginary ferret?” “I load the presents into the wheelbarrow.” Bill looked toward the driveway, where a battered wheelbarrow leaned beside a shed. “That’s your transport?” “Stealthy, spacious, environmentally responsible.” “It squeaks.” “Only if you push it.” “That is generally how wheelbarrows work.” Molly sighed. “Bill, criminal excellence requires imagination.” “Criminal excellence usually requires a van.” “Vans attract attention.” “So does a woman wheeling thirty-two presents down a suburban street.” “We’ll put a blanket over them.” “A blanket makes it worse. Then it’s obviously stolen presents.” A little girl in a silver crown appeared at the fence, staring directly at them. Molly froze. Bill froze harder. The girl narrowed her eyes. “Are you spies?” Bill smiled weakly. “No.” Molly smiled brighter. “Yes.” Bill turned to her in horror. The girl gasped. “Really?” “Top secret,” Molly said. “We’re investigating a dangerous criminal.” “Who?” Molly glanced at Bill. “A ferret thief.” Bill stared at her, then nodded solemnly. “Very dangerous. Possibly wearing glitter.” The girl considered this. Then she pointed toward the gazebo. “Sparklebeard has glitter.” Molly rose smoothly from behind the hedge. “Excellent work, Agent Princess. We’ll take it from here.” Bill followed, clutching the map. “Are we still stealing the presents?” Molly opened the side gate. “Of course.” “But now the birthday girl thinks we’re spies.” “Then let’s not disappoint her.”
created_at
"2026-07-09T17:54:26.204Z"
7/9/2026, 5:54:26 PM
model
"openai/gpt-5.6-terra-20260709"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
616
moderation_latency
(null)
generation_time
12581
tokens_prompt
77
tokens_completion
1298
native_tokens_prompt
66
native_tokens_completion
950
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
0.014415
router
(null)
provider_responses
0
endpoint_id
"3cc89000-ae34-4dde-9c7d-5ce009c8af0b"
id
"resp_0ec519979e40d4f5016a4fe0524c2c81908fadb8612de454ac"
is_byok
false
latency
273
model_permaslug
"openai/gpt-5.6-terra-20260709"
provider_name
"OpenAI"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.18.0; linux; x64))"
http_referer
(null)
request_id
"req-1783619666-vFPwDNhrCBuCMpLzv5DF"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1783619666-wUPh5KxCRwxF1P3cgSkH"
upstream_id
"resp_0ec519979e40d4f5016a4fe0524c2c81908fadb8612de454ac"
total_cost
0.014415
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
response_cache_source_id
(null)
data_region
"global"
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
684
0.00%
Dialogue to Total Word Ratio
Ratio: 60.32%, Deviation: 30.32%
neededClean
false
wordsTotal
688
wordsDialogue
415
0.0000%