NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-09-01KKACA2P0MSNRAMZX8DE680DN
Run Details
Model:
GPT-5.4 (Reasoning, Low)
Test:
Dialogue tags
Scenario:
Write 200 words with 10% dialogue
Duration:
30.5s
Cost:
$0.032970
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
100.00%
Dialogue to Total Word Ratio
(details)
100.0000%
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 surprised rabbit. Balloons bobbed over the garden fence; a hired magician was unloading rabbits with the grave expression of a bomb technician. On the patio, mountains of wrapped presents glittered beside a pink castle cake large enough to require planning permission. "Forty boxes," Molly whispered. "Fifty, counting the giant dinosaur. One squeaks." Bill swallowed. "Then we move fast, smile kindly, and fear glitter first, always." He adjusted his black turtleneck, which in daylight made him look less like a master thief and more like a depressed mime. Molly unfolded the blueprint: back gate, trampoline, juice table, and, at the center, the gift pile, guarded only by six-year-olds with sugar in their veins and whistles in their pockets. Their timetable was elegant. During musical statues, Bill would crawl under the trestle table and cut the ribbon netting. During pass-the-parcel, Molly would roll in the emergency wheelie bin, line it with a pirate tablecloth, and sweep the loot inside. Exit route: through the bouncy castle, over the paddling pool, into the alley before anyone noticed two extra adults. Inside, children sang off-key, and Bill looked ready to confess everything immediately.
created_at
"2026-03-09T22:43:29.097Z"
3/9/2026, 10:43:29 PM
model
"openai/gpt-5.4-20260305"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
28289
moderation_latency
(null)
generation_time
30182
tokens_prompt
77
tokens_completion
915
native_tokens_prompt
66
native_tokens_completion
2187
native_tokens_completion_images
0
native_tokens_reasoning
1917
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
0.03297
router
(null)
provider_responses
0
endpoint_id
"9ff5625c-403f-4d7f-b895-58ac7295062c"
id
"resp_0db9300d604cd2570169af4d1137ac81908dfb4f8897e22403"
is_byok
false
latency
259
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-1773096209-gq4jf3e0gsRTa36WfuTy"
upstream_id
"resp_0db9300d604cd2570169af4d1137ac81908dfb4f8897e22403"
total_cost
0.03297
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
100.00%
Dialogue to Total Word Ratio
Ratio: 9.76%, Deviation: 0.24%
neededClean
false
wordsTotal
205
wordsDialogue
20
100.0000%