NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-09-01KKABVYEQC48DYZ0NVQ81GB6G
Run Details
Model:
GPT-5.4 (Reasoning, Low)
Test:
Dialogue tags
Scenario:
Write 200 words with 10% dialogue
Duration:
18.5s
Cost:
$0.017715
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
98.51%
Dialogue to Total Word Ratio
(details)
99.2553%
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 studied the suburban battlefield from behind a hedge shaped like a startled dolphin. Beyond it, balloons bobbed over the garden gate, a clown assembled animals with the grim focus of a tax auditor, and sugared children pinballed across the lawn. Bill lay beside her in a caterer’s jacket stolen from a man still chasing his van three streets away. He unfolded the party map, which was really a napkin bearing jelly stains and one confident arrow. “Bill, we nick presents, not hostages,” Molly whispered. “Fine. In, grab, out. Smile if photographed,” Bill muttered. The target table sagged under a mountain of wrapped boxes, each tagged in looping handwriting and guarded only by a grandmother with opera glasses and the instincts of a prison warden. Molly adjusted her paper hat disguise, which read HAPPY 6TH, and felt professionally insulted by the level of security. A magician was sawing a sponge rabbit in half; this, she noted, might prove useful. Bill produced a confetti cannon, a gravy ladle, and two juice boxes. At the birthday song’s first shriek, they vaulted the hedge, looking less like criminal masterminds than very determined substitute uncles. Nobody noticed the loot trolley was actually a stroller.
created_at
"2026-03-09T22:35:46.25Z"
3/9/2026, 10:35:46 PM
model
"openai/gpt-5.4-20260305"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
17854
moderation_latency
(null)
generation_time
18054
tokens_prompt
77
tokens_completion
679
native_tokens_prompt
66
native_tokens_completion
1170
native_tokens_completion_images
0
native_tokens_reasoning
906
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
0.017715
router
(null)
provider_responses
0
endpoint_id
"9ff5625c-403f-4d7f-b895-58ac7295062c"
id
"resp_07a1b4651d2ac8000169af4b4259a48190b16112aa3862252f"
is_byok
false
latency
208
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-1773095746-FsYGHCJ6F1bfno83uIlj"
upstream_id
"resp_07a1b4651d2ac8000169af4b4259a48190b16112aa3862252f"
total_cost
0.017715
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
98.51%
Dialogue to Total Word Ratio
Ratio: 6.50%, Deviation: 3.50%
neededClean
false
wordsTotal
200
wordsDialogue
13
99.2553%