NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-09-01KX41BDH1129KF0B0TND2JRJF
Run Details
Model:
GPT-5.6 Terra (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
31.9s
Cost:
$0.064260
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
85.86%
Dialogue to Total Word Ratio
(details)
92.9304%
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
At 2:17 on Saturday afternoon, Molly and Bill crouched behind a municipal hedge across from the Bellwether house, where eight sugar-fueled children were celebrating Milo Bellwether's seventh birthday. Molly had a clipboard, a black turtleneck, and the calm expression of a woman who had once stolen a diamond necklace while disguised as a coat rack. Bill wore camouflage, although it was printed with cartoon ducks, because the only dark clothing he owned had been borrowed by a funeral magician. “The children will be distracted by the magician at two-thirty. He makes a rabbit disappear and looks under chairs. That is our window.” Inside, a pink balloon arch guarded the front door. Beyond it, parents chatted with paper cups in hand, displaying the universal exhaustion of adults trapped near a hired entertainer. “I dislike windows. Windows mean climbing, and climbing means knees. Why can't we enter normally, wearing paper hats and carrying casserole?” A man in a purple cape stumbled past the window, dragging a reluctant white rabbit and apologizing to it. The rabbit appeared to be winning the argument. “Because normal guests bring presents, Bill. We are removing them. Also, your casserole contains lock picks, three sausages, and a ferret.” At the curb sat their escape vehicle: a laundromat cart fitted with bicycle wheels, a brass bell, and a cardboard sign reading FRESH LINENS. Bill had painted the sign himself, then added a comma after “fresh” because he believed punctuation reassured witnesses. “Professor Mustard is a trained professional. He knows when a situation requires teeth.” Molly opened the clipboard. Its page contained a hand-drawn map, a timetable, and a list titled “Possible Obstacles,” beneath which Bill had written: “children, clowns, moral consequences.” She crossed out the last item with a firm, professional stroke. “The back gate is unlatched. We cross the garden, pass the inflatable pirate ship, and empty the gift table into the laundry cart. No teeth, no sausages, no improvising.” From the garden came a shriek, followed by applause. The magician had apparently produced a pigeon from a birthday cake. Several children chased it through the flowerbeds, while one solemn girl used a plastic telescope to inspect the sky for additional pastry birds. “What about gratitude? Children receive gifts with gratitude. Should we leave a note? Perhaps something educational, like, ‘Thank you for the drones; your future is in accounting.’” Bill's duck camouflage rustled as he lifted Professor Mustard from the casserole dish. The ferret wore a tiny burglar's mask made from an athletic sock and looked deeply offended by the assignment. “No notes. Last time you signed one ‘The Gentleman Crumb,’ and the police spent a week interviewing bakers.” Before Bill could answer, the front door opened. A small boy in a gold crown stood on the porch, holding a foam sword and staring directly at them. Behind him, the magician's rabbit bounded through the hall with a wrapped robot clamped beneath each ear. The boy pointed gardenward. Molly briskly took Bill's elbow and steered him inside, grinning.
created_at
"2026-07-09T18:13:30.279Z"
7/9/2026, 6:13:30 PM
model
"openai/gpt-5.6-terra-20260709"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2994
moderation_latency
(null)
generation_time
31304
tokens_prompt
77
tokens_completion
2006
native_tokens_prompt
66
native_tokens_completion
4273
native_tokens_completion_images
0
native_tokens_reasoning
3624
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
0.06426
router
(null)
provider_responses
0
endpoint_id
"3cc89000-ae34-4dde-9c7d-5ce009c8af0b"
id
"resp_0f56b6acfbed8b95016a4fe4ca56e08195ac73f7c6d71d6a7e"
is_byok
false
latency
215
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-1783620810-DsqaBSz9usGx9k3RVfU2"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1783620810-CA4MyUPU6Rn2XxkG9ftm"
upstream_id
"resp_0f56b6acfbed8b95016a4fe4ca56e08195ac73f7c6d71d6a7e"
total_cost
0.06426
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
100.00%
Matches word count
n/a
neededClean
false
words
501
85.86%
Dialogue to Total Word Ratio
Ratio: 31.75%, Deviation: 1.75%
neededClean
false
wordsTotal
504
wordsDialogue
160
92.9304%