NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M36NHV2Z5X0BZRAM5Q6FGQF2
Run Details
Model:
GPT-6 Luna (Reasoning, Medium)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
53.6s
Cost:
$0.003540
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
100.00%
Dialogue to Total Word Ratio
(details)
100.0000%
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
Rain ticked against the bakery awning while Molly and Bill crouched beside a suspiciously cheerful birthday banner, studying the community hall’s glowing windows; Molly whispered. “We enter, lift every present, and vanish. No alarms, no witnesses, and absolutely no accepting cake, because crumbs compromise any getaway plan on rainy nights.” Bill adjusted his paper crown, which read BIRTHDAY GUEST in glittering letters, though he had borrowed it from a scarecrow outside the garden center yesterday. “I object to calling cake a compromise. Cake is a morale tool, and morale is essential when confronting hostile balloons and tiny dancers after dark.” Molly unfolded their floor plan, drawn on a napkin and decorated with a smiling dinosaur; the hall’s layout had been guessed from its parking lot. “The presents sit under the table. We grab the shiny ones first, unless one contains socks; socks are morally indistinguishable from traps anyway, probably, ever.” From inside came a chorus of children singing off-key, followed by an adult’s attempt to conduct them using paper plates and an expression of surrender. “There are children. I assumed a party meant adults wearing hats. We should reconsider, preferably before someone asks me to play duck-duck-goose again, immediately, please.” Molly lowered the napkin. Her expression softened, then sharpened as she noticed a mountain of wrapped boxes beneath a banner announcing the birthday child’s name. “We are not stealing from children. We are relocating gifts temporarily, then returning them after the cake, when confusion becomes tradition for everyone eventually, anyway.” Bill considered this loophole, then glanced at the enormous toy rocket visible through the window, its cardboard fins trembling whenever the radiator coughed like applause. “Fine. We return everything. But if that rocket has a launch button, I reserve the right to inspect it for safety and quality quite carefully.” A balloon escaped the hall and bobbed above them, tugging its string like a tiny, judgmental moon; Bill saluted it, perhaps seeking professional approval there. “We could simply leave. That would be lawful, easy, and devastating to our brand. The neighborhood expects us to make terrible decisions together as professionals.” Bill opened his coat, revealing not burglary tools but a gift bag, tissue paper, and a card reading Sorry We Were Ridiculous, written in letters. “Plan revised. We sneak in, apologize, and offer to help wrap presents. If they refuse, we escape with dignity and possibly frosting anyway right now.” They slipped through the open side door and entered beneath a paper-chain arch, where a child in a pirate hat promptly blocked their path completely. “Are you the present inspectors? Because Uncle Greg ate the label on mine, and now nobody knows whether I’m getting dinosaurs today before the cake.” Molly and Bill exchanged a look, then lifted the gift bag. Behind them, the adults applauded, mistaking their guilty expressions for enthusiastic volunteerism at last. “We can inspect the labels. And retrieve Uncle Greg’s lunch if necessary. Nobody steals presents tonight; we steal the show, apparently, and file a report.”
created_at
"2026-09-23T08:19:12.101Z"
9/23/2026, 8:19:12 AM
model
"openai/gpt-6-luna-20260922"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3266
moderation_latency
(null)
generation_time
53518
tokens_prompt
75
tokens_completion
1342
native_tokens_prompt
66
native_tokens_completion
7067
native_tokens_completion_images
0
native_tokens_reasoning
6433
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
0.0035401
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.20.0; linux; x64))"
http_referer
(null)
request_id
"req-1790151552-gdEY2RJnaYzFnTIeYatC"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790151552-HuyCxXpPbJuxcAqH2EKi"
upstream_id
"resp_0dad2d44a42ac39b016ab38b80340087d1beb160dfccae8c9f"
provider_responses
0
endpoint_id
"05e94e02-b9c2-4bbb-ba55-4082ee9ad687"
id
"resp_0dad2d44a42ac39b016ab38b80340087d1beb160dfccae8c9f"
is_byok
false
latency
436
model_permaslug
"openai/gpt-6-luna-20260922"
provider_name
"OpenAI"
status
200
total_cost
0.0035401
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
100.00%
Matches word count
n/a
neededClean
false
words
500
100.00%
Dialogue to Total Word Ratio
Ratio: 50.00%, Deviation: 0.00%
neededClean
false
wordsTotal
504
wordsDialogue
252
100.0000%