NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-09-01KK9PYDEJ3W2FYMAMRZ7B46NV
Run Details
Model:
GPT-5.4 (Reasoning)
Test:
Dialogue tags
Scenario:
Write 200 words with 90% dialogue
Duration:
46.6s
Cost:
$0.044850
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
97.47%
Dialogue to Total Word Ratio
(details)
98.7362%
Overall
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
“Children’s parties are chaos with frosting. Nobody counts boxes when a clown is falling over.” “I hate clowns. They’re freelancers in makeup.” “Focus. We stroll in as entertainers.” “I can juggle anxiety and one orange.” “Perfect. I’ll do balloons.” “You can’t make a balloon dog.” “I can make a balloon disappointment. Parents love realism.” “What about the presents?” “We stack them by the door, call it a treasure hunt, then treasure leaves.” “That sounds illegal even by our standards.” “We’re thieves, Bill, not tax consultants.” “What if a child asks for a rabbit from my hat?” “Give them a lecture on supply chains.” “What if there’s a bouncy castle?” “Then we bounce with dignity.” “I have never bounced with dignity.” “You don’t need dignity. You need a smile and shoes that squeak.” “My shoes honk.” “Better. Memorable branding.” “And the cake?” “Untouched. We’re crooks, not savages.” Bill sighed. “I feel odd robbing a room full of tiny monarchs.” “Think of it as wealth redistribution from people too young to file complaints.” “Smile, Bill. Nothing disarms suspicion like confidence and a ridiculous tie.” From hallway came shrieking, music, and glitter. Bill adjusted tie. “If I’m arrested, tell them I’m a magician.”
created_at
"2026-03-09T16:30:06.812Z"
3/9/2026, 4:30:06 PM
model
"openai/gpt-5.4-20260305"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
43622
moderation_latency
(null)
generation_time
46488
tokens_prompt
77
tokens_completion
1166
native_tokens_prompt
66
native_tokens_completion
2979
native_tokens_completion_images
0
native_tokens_reasoning
2692
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
0.04485
router
(null)
provider_responses
0
endpoint_id
"9ff5625c-403f-4d7f-b895-58ac7295062c"
id
"resp_02e9166460ad4ab20169aef58ee5948193a7f4b61d92be67fb"
is_byok
false
latency
279
model_permaslug
"openai/gpt-5.4-20260305"
provider_name
"OpenAI"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.0; linux; x64))"
http_referer
(null)
api_type
"completions"
id
"gen-1773073806-idwjJJht9g7P6XdgA5nT"
upstream_id
"resp_02e9166460ad4ab20169aef58ee5948193a7f4b61d92be67fb"
total_cost
0.04485
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
Evaluation details
Result
Evaluator
Details
Meta Data
100.00%
Matches word count
n/a
neededClean
false
words
200
97.47%
Dialogue to Total Word Ratio
Ratio: 94.00%, Deviation: 4.00%
neededClean
false
wordsTotal
200
wordsDialogue
188
98.7362%