NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-16-01KHKAEEF9J71BYH5TRZ551SBD
Run Details
Model:
GPT-5
Test:
Dialogue tags
Scenario:
Write 200 words with 90% dialogue
Duration:
1m 26s
Cost:
$0.051682
Provider:
OpenAI
Model input
User/Human
You are a author of comedic heist books. Please write a 200 word scene that has 90% 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: I cannot believe our biggest score involves glitter and sheet cake. Bill: Crime has to evolve; I'm pivoting to party logistics. Molly: We're not planners, Bill. We're pirates, but with better shoes. Bill: Then consider this a treasure hunt, and the X is a bouncy castle. Molly: The castle squeaks. Subtlety squeaks back. Bill: Distraction first. I do a tragic balloon giraffe; hearts melt; gift table loosens. Molly: Your last balloon animal looked like a question mark with allergies. Bill: Mystery breeds generosity. Molly: And suspicion. Bill: Fine. You charm the toddlers; I'll wrangle the wrapping paper. Molly: Toddlers can smell fear and raisins. Bill: I smell like neither. I exude wholesome chaos. Molly: The parents? Bill: We blend—smile, nod, compliment the cake density. Molly: Complimenting density sounds like flirting with granite. Bill: Granite is dependable. Molly: Presents are guarded by an aunt with elbows like traffic cones. Bill: I respect elbows. I also respect decoys. We move decoy boxes; real ones follow. Molly: You're proposing reverse Santa, but with worse cardio. Bill: Accurate. Also, I brought name tags. I'm "Uncle Who." Molly: I'm "Cousin Probably." Bill: Confidence is key. Never run; just stroll with purpose and streamers. Molly: And if karaoke starts? Bill: We duet, bow, and exit with harmonized parcels. Molly: Bold. Petty. Festive. Let's mourn our dignity afterward. Silence.
created_at
"2026-02-16T13:34:08.148712+00:00"
2/16/2026, 1:34:08 PM
model
"openai/gpt-5-2025-08-07"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
11865
moderation_latency
(null)
generation_time
86193
tokens_prompt
77
tokens_completion
1156
native_tokens_prompt
66
native_tokens_completion
5160
native_tokens_completion_images
0
native_tokens_reasoning
4800
native_tokens_cached
0
num_media_prompt
(null)
num_input_audio_prompt
(null)
num_media_completion
0
num_search_results
(null)
origin
"https://ncbench.com/"
is_byok
false
finish_reason
"stop"
native_finish_reason
"completed"
usage
0.0516825
router
(null)
provider_responses
0
id
"resp_0bcb80350d7724f00169931c7a97888194bd7613f15ad6eb2c"
status
200
is_byok
false
latency
1393
endpoint_id
"7c2f859a-7890-4e8e-b1de-1cd1c0a800b4"
provider_name
"OpenAI"
model_permaslug
"openai/gpt-5-2025-08-07"
api_type
"completions"
id
"gen-1771248761-sexktdpnXogSqW7eRuTO"
upstream_id
"resp_0bcb80350d7724f00169931c7a97888194bd7613f15ad6eb2c"
total_cost
0.0516825
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
Evaluation details
Result
Evaluator
Details
Meta Data
3.6234%
Matches word count
n/a
neededClean
false
words
224
0.0000%
Dialogue to Total Word Ratio
Ratio: 1.79%, Deviation: 88.21%
neededClean
false
wordsTotal
224
wordsDialogue
4
1.8117%