NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-09-01KX414KJXPCN4408QWHSE5F6P
Run Details
Model:
GPT-5.6 Terra
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
13.6s
Cost:
$0.014505
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
4.30%
Dialogue to Total Word Ratio
(details)
2.1514%
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
“Children’s party,” Bill whispered, peering through the hedge. “I still say it’s beneath us.” “Nothing is beneath us,” Molly said. “We once stole a mayor’s ceremonial spade.” “It was gold-plated.” “It was spray-painted.” “It had emotional value.” “So do these presents.” Bill adjusted his black knit cap. “You’re certain they’re presents?” “Wrapped boxes, bows, tiny bicycles, one suspiciously expensive-looking telescope. Unless children have begun giving one another home appliances, yes.” “Children are unpredictable. I had a nephew who asked for a vacuum cleaner.” “That explains why you don’t visit him.” “He was seven. He knew too much.” Molly unfolded a crayon-covered map. “The house belongs to the Pembertons. Party in the garden. Clown at two o’clock. Cake at three. Gift table under the gazebo.” “Any dogs?” “One.” “What breed?” “Labrador.” Bill relaxed. “Excellent. I can negotiate with a Labrador.” “You cannot negotiate with a Labrador.” “I once talked one into giving me half a sandwich.” “You dropped half a sandwich.” “It understood the terms.” A shriek erupted from the garden, followed by the sound of a whistle and a chorus of children chanting, “Again! Again!” Bill flinched. “What is that?” “Party games.” “Sounds like civil unrest.” “It’s musical statues.” “Ah. The cruelest game of all. Teaching children that movement is punishable.” Molly pointed to the side gate. “We go in there. You distract the clown.” “I hate clowns.” “Everyone hates clowns. That’s why it’ll work.” “What do I say?” “Ask him where he keeps the balloon pump.” “He’ll know it’s a trick.” “Then ask him something clownish.” Bill thought. “Do your shoes come in sensible?” “Perfect. He’ll be furious.” “And you?” “I slip behind the gazebo, load the presents into the laundry trolley, and wheel them out.” “What laundry trolley?” “The one I stole from the neighbour’s washing line.” Bill stared at her. “You stole a trolley from a washing line?” “It was unattended.” “It’s a basket on wheels.” “It’s transport infrastructure.” A child’s voice called, “Mum! The magician turned Dad into a rabbit!” Bill squinted through the hedge. “There’s a magician too?” “Potential complication.” “Potential professional jealousy.” “We are not discussing your failed career in stage magic.” “My career was sabotaged by a hostile dove.” “You released it indoors.” “It had stage fright.” Molly tucked the map away. “Focus. At precisely two fifteen, the children begin the treasure hunt.” “Why does that matter?” “Because every adult will be following tiny pirates around the garden while shouting, ‘Warmer! Colder!’” Bill nodded slowly. “That is diabolical.” “It’s parenting.” “What if the presents are labelled?” “They will be.” “Then everyone will know whose present is missing.” “Not if we replace them.” Bill blinked. “Replace them with what?” Molly opened her coat. Inside were several bundles wrapped in newspaper and glittery paper. “Cabbages?” Bill asked. “Decorated cabbages.” “One has a bow.” “They all have bows.” “Why cabbages?” “Children rarely inspect presents before opening them. They see wrapping paper, they scream, they tear, they discover vegetables. By then we’re gone.” Bill looked impressed despite himself. “That’s monstrous.” “Thank you.” “And the telescope?” “Sold to a collector.” “The bicycles?” “Sold individually.” “The talking dinosaur?” Molly paused. “We keep the dinosaur.” “Professional standards.” “Exactly.” The side gate creaked open. A Labrador waddled toward them, tail wagging. Bill crouched. “Hello, officer. Lovely evening for crime.” The dog licked his face. Molly sighed. “Negotiating?” “He’s agreed to look the other way for sausages.” “You brought sausages?” “I plan ahead.” From the garden, the clown shouted, “Who wants a balloon sword?” Every child screamed. Bill rose, pale. “Right. I distract the nightmare.” Molly grabbed the trolley. “And I steal Christmas, birthdays, and possibly a small blender.” “Two fifteen?” “Two fifteen.” Bill marched toward the gate, then stopped. “Molly?” “Yes?” “If there’s a cabbage with my name on it…” “There won’t be.” “Good.” “It’ll have your mother’s name.” Bill grinned. “Now that’s a gift.”
created_at
"2026-07-09T18:09:47.122Z"
7/9/2026, 6:09:47 PM
model
"openai/gpt-5.6-terra-20260709"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
635
moderation_latency
(null)
generation_time
13291
tokens_prompt
77
tokens_completion
1379
native_tokens_prompt
66
native_tokens_completion
956
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
0.014505
router
(null)
provider_responses
0
endpoint_id
"3cc89000-ae34-4dde-9c7d-5ce009c8af0b"
id
"resp_0d97545cf39b5e7c016a4fe3eb59b081908a8afd1fe665e30d"
is_byok
false
latency
286
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-1783620587-VJxe1BQyWBuaGrGbUm9Y"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1783620587-0qSiJVN2wQXnIwi0Ry1U"
upstream_id
"resp_0d97545cf39b5e7c016a4fe3eb59b081908a8afd1fe665e30d"
total_cost
0.014505
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
644
4.30%
Dialogue to Total Word Ratio
Ratio: 77.93%, Deviation: 7.93%
neededClean
false
wordsTotal
648
wordsDialogue
505
2.1514%