NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-09-01KK9QJQJEYP3NPAQKFQ96MHY8
Run Details
Model:
GPT-5.4 (Reasoning)
Test:
Dialogue tags
Scenario:
Write 200 words with 90% dialogue
Duration:
11.9s
Cost:
$0.006735
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
43.38%
Matches word count
(details)
36.79%
Dialogue to Total Word Ratio
(details)
40.0831%
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, you cannot wear the striped jumper.” “Why not? It says ‘harmless uncle.’” “It says ‘escaped mime.’ We need subtle.” “Molly, we’re robbing a castle made of balloons.” “A children’s party, Bill, not a bank. That makes subtle more important.” “I still say the clown is our way in.” “No clowning. Clowns are remembered, feared, and occasionally bitten.” “Fine. We go in as caterers.” “With what food?” “You’re holding six sausage rolls.” “I’m holding our lunch.” “Then we go in as magicians.” “You know one trick.” “It’s a good trick.” “It’s stealing watches from pensioners.” “Transferable skill.” “Bill, the objective is the presents, not a public inquiry.” “Right. Presents. We sweep the gift table, out the side gate, into the van.” “The van with ‘Definitely Plumbing’ painted on it?” “It discourages questions.” “It encourages them. Why ‘definitely’?” “Confidence.” “Idiocy. What about the children?” “What about them?” “They’ll notice two strangers lugging twenty-seven boxes and a bicycle.” “Then we distract them with cake.” “You cannot solve everything with cake.” “I can solve most things with cake.” “Listen: smile, wave, say we’re from the magician, locate the presents, and leave quietly.” “And if someone asks where the magician is?” “Say he’s stuck in a hat.” “Molly, that’s absurd.” “Bill, we’re stealing a pony-themed toaster from eight-year-olds. Absurd has won.”
created_at
"2026-03-09T16:41:12.536Z"
3/9/2026, 4:41:12 PM
model
"openai/gpt-5.4-20260305"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2499
moderation_latency
(null)
generation_time
11768
tokens_prompt
77
tokens_completion
466
native_tokens_prompt
66
native_tokens_completion
438
native_tokens_completion_images
0
native_tokens_reasoning
105
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
0.006735
router
(null)
provider_responses
0
endpoint_id
"9ff5625c-403f-4d7f-b895-58ac7295062c"
id
"resp_027fcf8881aa45de0169aef828a0788193ac65addffef0cdd6"
is_byok
false
latency
164
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-1773074472-RXFsYoyAocEdtngwwneI"
upstream_id
"resp_027fcf8881aa45de0169aef828a0788193ac65addffef0cdd6"
total_cost
0.006735
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
Evaluation details
Result
Evaluator
Details
Meta Data
43.38%
Matches word count
n/a
neededClean
false
words
217
36.79%
Dialogue to Total Word Ratio
Ratio: 100.00%, Deviation: 10.00%
neededClean
false
wordsTotal
221
wordsDialogue
221
40.0831%