NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-16-01KHK9PA0Z5Y0WKE8EE7722QN2
Run Details
Model:
GPT-5
Test:
Dialogue tags
Scenario:
Write 200 words with 90% dialogue
Duration:
1m 48s
Cost:
$0.038812
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
"Bill, you brought a spreadsheet to a birthday party?" "Molly, crime without columns is just chaos." "We're not stealing chaos; we're liberating wrapping paper." "And the mysterious boxes within, which we will ethically relocate." "Ethically? To our apartment?" "Think of it as a foster home for neglected toys." "The toys are currently loved." "Yes, but inefficiently loved; we can optimize the affection supply chain." "So what's the plan that doesn't end with frosting in my hair?" "We blend in, harness the ancient art of 'yay,' and migrate presents into this nondescript sack." "That's a unicorn backpack." "Exactly, aggressively descript." "What about cake guardians and balloon bouncers?" "We deploy diversion: I perform my acclaimed invisible magic show." "You stand still and claim things vanished." "It killed at my cousin's thanksgiving. The turkey never recovered." "Meanwhile, I what, juggle excuses?" "Yes. Say you're the Gift Auditor. Stamp things. People love stamps." "I have a potato." "Even better. Rustic credibility." "And when someone asks whose party we're here for?" "Easy: we cheer 'All of them!' Inclusivity confuses authority." "I can't believe I'm doing this in squeaky shoes." "Confidence squeaks, Molly." "And if a kid cries?" "We cry harder. Out-sad them. Works every time."
created_at
"2026-02-16T13:21:18.289884+00:00"
2/16/2026, 1:21:18 PM
model
"openai/gpt-5-2025-08-07"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
9126
moderation_latency
(null)
generation_time
107410
tokens_prompt
77
tokens_completion
1113
native_tokens_prompt
66
native_tokens_completion
3873
native_tokens_completion_images
0
native_tokens_reasoning
3584
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
0.0388125
router
(null)
provider_responses
0
id
"resp_02fa25d3d2f589ec01699319628e108197b071f984be726a57"
status
200
is_byok
false
latency
284
endpoint_id
"7c2f859a-7890-4e8e-b1de-1cd1c0a800b4"
provider_name
"OpenAI"
model_permaslug
"openai/gpt-5-2025-08-07"
api_type
"completions"
id
"gen-1771247970-OGv8NmIY6fNvTLriTYoj"
upstream_id
"resp_02fa25d3d2f589ec01699319628e108197b071f984be726a57"
total_cost
0.0388125
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
Evaluation details
Result
Evaluator
Details
Meta Data
99.9990%
Matches word count
n/a
neededClean
false
words
201
36.7879%
Dialogue to Total Word Ratio
Ratio: 100.00%, Deviation: 10.00%
neededClean
false
wordsTotal
201
wordsDialogue
201
68.3935%