NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-09-01KKABZXEBJZKK551VDE4ZF0XN
Run Details
Model:
GPT-5.4 (Reasoning, Low)
Test:
Dialogue tags
Scenario:
Write 200 words with 50% dialogue
Duration:
10.9s
Cost:
$0.005490
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
35.00%
Matches word count
(details)
99.98%
Dialogue to Total Word Ratio
(details)
67.4926%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 200 word scene that has 50% 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, studying the bungalow as if it were a mint and not a house with balloon arches and a cardboard dragon taped to the door. Through the window, children in paper crowns swarmed around a mountain of presents. “Bill, this is either genius or the moment we admit we’ve lost all professional standards.” “Standards are for thieves with pensions,” Bill whispered, adjusting a tiny plastic tiara he’d found in the gutter. “Observe the weaknesses: one clown, two exhausted parents, and a piñata distracting the entire security team.” Molly unfolded a crayon-scribbled floor plan. It smelled faintly of jam. “We go in as entertainers, sweep the loot into the laundry basket, and exit during cake deployment.” “Bold. Elegant. Slightly sticky. I love it.” Inside the gate, a magician was already bombing. A rabbit had escaped and was eating crisps under a chair. “Change of plan,” Molly said. “Chaos has opened a window.” Bill peered over the hedge. “Literally. The kitchen window’s open.” They slipped across the lawn, keeping low behind a bouncy castle shaped like a pirate ship. “If anyone asks,” Molly murmured, “you’re Uncle Pete.” “I can’t be Uncle Pete. I look like a man banned from aquariums.” “Fine. You’re the balloon emergency technician.” Bill straightened. “At last, a title worthy of my talents.”
created_at
"2026-03-09T22:37:56.053Z"
3/9/2026, 10:37:56 PM
model
"openai/gpt-5.4-20260305"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3514
moderation_latency
(null)
generation_time
10814
tokens_prompt
77
tokens_completion
490
native_tokens_prompt
66
native_tokens_completion
355
native_tokens_completion_images
0
native_tokens_reasoning
49
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
0.00549
router
(null)
provider_responses
0
endpoint_id
"9ff5625c-403f-4d7f-b895-58ac7295062c"
id
"resp_0286dd4579b86a590169af4bc41f788197a2803836f397a13d"
is_byok
false
latency
166
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-1773095876-wyXnvjfmcQ9YndAbz0Tv"
upstream_id
"resp_0286dd4579b86a590169af4bc41f788197a2803836f397a13d"
total_cost
0.00549
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
Evaluation details
Result
Evaluator
Details
Meta Data
35.00%
Matches word count
n/a
neededClean
false
words
218
99.98%
Dialogue to Total Word Ratio
Ratio: 48.86%, Deviation: 1.14%
neededClean
false
wordsTotal
219
wordsDialogue
107
67.4926%