NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-09-01KK9QBGY8WBAG3GTJGJX0Z7DX
Run Details
Model:
GPT-5.4 (Reasoning)
Test:
Dialogue tags
Scenario:
Write 200 words with 10% dialogue
Duration:
29.1s
Cost:
$0.028860
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
100.00%
Dialogue to Total Word Ratio
(details)
99.9997%
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 garden hedge, studying the rented bouncy castle swelling like a smug pink lung. Beyond it, a suburban lawn glittered with juice boxes, paper crowns, and a mountain of presents under a banner that announced HUGO IS FIVE AND ARMED WITH CAKE. Bill lay beside her, wearing black gloves and the expression of a man trying to outwit primary colors. They had cased banks, museums, and once a suspiciously fortified pet shop, but nothing had prepared them for the security nightmare of twelve sugared children orbiting a hired princess and a barbecue. Every ten seconds, another parent emerged from the house carrying snacks and the exhausted smile of someone one balloon away from mutiny. Molly unfolded the plan on a napkin stained with mustard and dread. The route involved the side gate, the gift table, and a tactical retreat through the inflatable dinosaur tunnel. Bill adjusted the red foam nose he had bought for disguise and dignity. “Why clown suits?” Molly whispered. “Because nobody suspects a balloon giraffe specialist.” “And the toddlers?” “We blame the magician, moonwalk out, smiling, sweetly.” The magician sneezed fire into punch, and both thieves immediately promoted him from suspect to useful distraction.
created_at
"2026-03-09T16:37:16.37Z"
3/9/2026, 4:37:16 PM
model
"openai/gpt-5.4-20260305"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
27095
moderation_latency
(null)
generation_time
29029
tokens_prompt
77
tokens_completion
330
native_tokens_prompt
66
native_tokens_completion
1913
native_tokens_completion_images
0
native_tokens_reasoning
1655
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
0.02886
router
(null)
provider_responses
0
endpoint_id
"9ff5625c-403f-4d7f-b895-58ac7295062c"
id
"resp_04a338b669f066440169aef73c7068819595e6aae3edddc2e4"
is_byok
false
latency
214
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-1773074236-35BiKp13w15ztlf7pUkb"
upstream_id
"resp_04a338b669f066440169aef73c7068819595e6aae3edddc2e4"
total_cost
0.02886
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: 10.50%, Deviation: 0.50%
neededClean
false
wordsTotal
200
wordsDialogue
21
99.9997%