NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-09-01KX418J00F8SXYHTQ5TPDWK4T
Run Details
Model:
GPT-5.6 Terra
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
13.0s
Cost:
$0.014160
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
45.98%
Dialogue to Total Word Ratio
(details)
22.9884%
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
“Tell me again why we’re robbing a six-year-old,” Molly whispered. “We’re not robbing a six-year-old,” Bill whispered back. “We’re redistributing luxury goods from a chaotic domestic marketplace.” “It’s a birthday party.” “Exactly. Nobody audits a birthday party. Parents are distracted, children are sticky, and there’s always one uncle asleep in a deck chair.” “You’ve planned this before.” “I’ve attended them before. There’s a difference.” Molly peered through the hedge at the bungalow garden. Balloons bobbed over a fence. A pirate flag drooped beside a plastic paddling pool. Twenty children shrieked around a man in a foam parrot costume. “Those presents are wrapped in dinosaur paper,” she said. “That feels morally complicated.” “Morality is a luxury of people who don’t need a new toaster.” “We stole a toaster last week.” “It was a disappointing toaster.” “You said it had eight settings.” “It had seven settings and a button marked ‘bagel.’ That’s not sophistication, that’s arrogance.” A child raced past the hedge carrying a foam sword. “Look at them,” Molly said. “They’re tiny.” “Tiny people can own expensive things. There’s a remote-control drone in that pile. I saw the box.” “For the birthday child.” “Probably. But perhaps he’s irresponsible.” “He’s six.” “Then statistically, extremely irresponsible.” Molly adjusted her black knit cap. “What’s the plan?” Bill unfolded a crumpled map drawn on the back of a takeaway menu. “Phase one: enter through the side gate during the magician’s finale.” “There’s a magician?” “There is always a magician. They are the natural predator of adult attention.” “And phase two?” “You pretend to be a children’s entertainer.” “I don’t have a costume.” Bill handed her a plastic eye patch and a cardboard hook. “This is your costume?” “You’re Pirate Molly.” “I’m wearing trousers and a burglary jacket.” “Pirates were practical.” “What are you?” “Balloon technician.” “That’s not a thing.” “It is today.” Bill lifted a portable pump from his bag. “I inflate three balloons, complain about helium prices, and no one questions me.” “That might actually work.” “Thank you.” “I said might.” “You wound me.” From the garden came a loud cry: “Abraca-sausages!” Bill nodded grimly. “The magician is entering his peak nonsense. This is our window.” Molly grabbed his sleeve. “Wait. What do we do with the presents after?” “We take them to the van.” “And then?” “Sort them by resale value.” “They’re children’s presents, Bill. Most of them will be slime.” “Slime has a market.” “No, it has a smell.” “A very popular smell, apparently.” They slipped through the side gate. Bill immediately began pumping a yellow balloon. A mother in a sunhat frowned at him. “Who are you?” Bill didn’t miss a stroke. “Balloon safety inspection.” “Oh. Are they safe?” “Not emotionally,” he said. “But structurally, excellent.” She nodded and hurried away. Molly stared at him. “You’re horrifyingly good at that.” “People believe confidence. It’s the only free resource left.” A boy wearing a paper crown marched up to Molly. “Are you a pirate?” “Absolutely,” she said. “Where’s your ship?” “Parked illegally around the corner.” “Can you fight the parrot man?” Molly looked toward the foam parrot, who was attempting to juggle oranges and losing an argument with gravity. “I could,” she said. “But today I am on a peace mission.” The boy pointed at the presents. “Those are mine.” “Of course they are.” “You better not steal them.” Molly froze. Bill coughed loudly. “Children say the funniest things.” The boy narrowed his eyes. “I have a dog.” “Wonderful,” Molly said. “He bites pirates.” “Then I’m actually a balloon technician.” Bill hissed, “No, you’re not.” “I’m diversifying.” The magician shouted, “And now, children, watch as I make this entire pile of gifts disappear!” Every adult turned toward him. Bill looked at Molly. “Well?” Molly glanced at the presents, then at the birthday boy’s suspicious little face. “New plan,” she said. “What new plan?” “We steal the magician’s wallet.” Bill smiled. “Now you’re thinking professionally.”
created_at
"2026-07-09T18:11:56.556Z"
7/9/2026, 6:11:56 PM
model
"openai/gpt-5.6-terra-20260709"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
622
moderation_latency
(null)
generation_time
12800
tokens_prompt
77
tokens_completion
1313
native_tokens_prompt
66
native_tokens_completion
933
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
0.01416
router
(null)
provider_responses
0
endpoint_id
"3cc89000-ae34-4dde-9c7d-5ce009c8af0b"
id
"resp_0c01c52be8591267016a4fe46cad1c81909586733b7176d500"
is_byok
false
latency
324
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-1783620716-n5E4wRwyS8A0lWeyQeQZ"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1783620716-TmT8F8ZORCk3HUEBZpLL"
upstream_id
"resp_0c01c52be8591267016a4fe46cad1c81909586733b7176d500"
total_cost
0.01416
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
652
45.98%
Dialogue to Total Word Ratio
Ratio: 66.06%, Deviation: 3.94%
neededClean
false
wordsTotal
657
wordsDialogue
434
22.9884%