NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-18-01KM08TPSBAQW0EHFWYE2W7HEV
Run Details
Model:
GPT-5.4 Nano
Test:
Dialogue tags
Scenario:
Write 200 words with 10% dialogue
Duration:
2.9s
Cost:
$0.000434
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.03%
Dialogue to Total Word Ratio
(details)
0.0164%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 200 word scene that has 10% 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 crouched behind the hedge, patting her satchel like it contained a carefully folded blanket instead of crowbars and guilt. Bill stood at the curb with a handheld scanner the size of a taco, blinking as if it too was nervous about cake. “Okay,” Molly whispered, staring at the front window where balloons bobbed like cheerful sentries, “that house is basically a candy-coated fortress.” Bill lifted a finger toward a tiny motion sensor mounted beside the porch swing. “Relax. I’ve got a plan. We’ll charm the sensor with science.” He tapped the scanner. It emitted a polite chirp, then immediately squealed, as though offended by his confidence. Molly leaned closer to the door. “We are absolutely not charming anything. We are sneaking. Like professionals.” She slid a wafer-thin wedge into the lock with the delicacy of someone cutting birthday cake, except with more consequences. The latch clicked; the door swung inward just enough to invite trouble. From inside came laughter, the kind that sounded like sprinkles and shushing. Bill peered through the crack. “If there’s a guard dog, I’m naming it first.” Molly smirked. “If there’s a guard dog, you’re feeding it the bribe paperwork.” They slipped into the hallway, where every gift sat under a towering tree of wrapping paper. Molly’s eyes gleamed. “Ready to snatch everything?” she asked. Bill grinned. “After you promise not to steal the dinosaur piñata.” Molly nodded. “No promises.” Then they began, quietly, eagerly, and very badly timed.
created_at
"2026-03-18T10:45:57.176Z"
3/18/2026, 10:45:57 AM
model
"openai/gpt-5.4-nano-20260317"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1432
moderation_latency
(null)
generation_time
2746
tokens_prompt
77
tokens_completion
410
native_tokens_prompt
66
native_tokens_completion
337
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
0.00043445
router
(null)
provider_responses
0
endpoint_id
"0c835f2e-c18d-4e8c-b245-e1e3bd08b97f"
id
"resp_0cbb69d896c071be0169ba82653b248196b6d445e663410a06"
is_byok
false
latency
1171
model_permaslug
"openai/gpt-5.4-nano-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-1773830757-i0QxFGC7jjmJ2SEYAvaR"
upstream_id
"resp_0cbb69d896c071be0169ba82653b248196b6d445e663410a06"
total_cost
0.00043445
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
244
0.03%
Dialogue to Total Word Ratio
Ratio: 26.83%, Deviation: 16.83%
neededClean
false
wordsTotal
246
wordsDialogue
66
0.0164%