NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-18-01KM08PKH9JD89KTPB4KD1GXYM
Run Details
Model:
GPT-5.4 Mini
Test:
Dialogue tags
Scenario:
Write 200 words with 90% dialogue
Duration:
2.4s
Cost:
$0.001539
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
0.21%
Matches word count
(details)
92.13%
Dialogue to Total Word Ratio
(details)
46.1716%
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
“Bill, are you sure this is the right party?” Molly whispered. “Unless the banner says ‘Happy Birthday, Timmy,’ in glowing balloons, yes,” Bill said. “It does say that.” “Then we are blessed with excellent intelligence.” “And terrible morals.” “Same team, Molly.” Molly adjusted her black gloves. “You said the presents would be in the garage.” “They will be. That is where people put loot when they have no security mindset.” “It’s a children’s party, Bill. The security mindset is sugar and supervision.” “Exactly. Adults are distracted by juice boxes and liability.” “I hate how confident you sound.” “I hate how right I usually am.” A burst of laughter floated from inside the house. Molly frowned. “What if we’re caught?” “We won’t be.” “That’s what you said before the petting zoo.” “That donkey had commitment.” “Bill.” “Fine. If we’re caught, we act like entertainers.” “You think the guards will believe that?” “It’s a children’s party. No one knows what’s going on. One minute a clown is twisting swords, the next minute he’s in a lawsuit.” Molly peeked through the hedge. “I see the gift table.” “Beautiful.” “It’s enormous.” “Even more beautiful.” “And there’s a child sleeping on it.” Bill paused. “That complicates things.” Molly sighed. “So we’re not stealing the presents.” Bill stared at the glittering pile. “Of course we are.” “We can’t.” “We can.” “Bill, no.” “Bill, yes.”
created_at
"2026-03-18T10:43:42.774Z"
3/18/2026, 10:43:42 AM
model
"openai/gpt-5.4-mini-20260317"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
378
moderation_latency
(null)
generation_time
1980
tokens_prompt
77
tokens_completion
464
native_tokens_prompt
66
native_tokens_completion
331
native_tokens_completion_images
0
native_tokens_reasoning
0
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
0.001539
router
(null)
provider_responses
0
endpoint_id
"9ee065b2-3d1c-43bc-bdd7-28af3b148282"
id
"resp_0928acc38cff2f1c0169ba81ded69c8193a43ce7e0a666bc65"
is_byok
false
latency
212
model_permaslug
"openai/gpt-5.4-mini-20260317"
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-1773830622-L2AeqENuSmQRfUBcTCKZ"
upstream_id
"resp_0928acc38cff2f1c0169ba81ded69c8193a43ce7e0a666bc65"
total_cost
0.001539
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
Evaluation details
Result
Evaluator
Details
Meta Data
0.21%
Matches word count
n/a
neededClean
false
words
228
92.13%
Dialogue to Total Word Ratio
Ratio: 84.65%, Deviation: 5.35%
neededClean
false
wordsTotal
228
wordsDialogue
193
46.1716%