NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-18-01KM09B65W0A6HFKR40ZVD1F3V
Run Details
Model:
GPT-5.4 Nano (Reasoning, Low)
Test:
Dialogue tags
Scenario:
Write 200 words with 90% dialogue
Duration:
13.4s
Cost:
$0.002746
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
3.62%
Matches word count
(details)
99.94%
Dialogue to Total Word Ratio
(details)
51.7824%
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
Molly: “Bill, tonight we infiltrate the children’s party.” Bill: “Relax, we’re professional. I brought my tiny crowbar and my larger confidence.” Molly: “Your confidence squeaks when you lie.” Bill: “It’s my stealth squeak. Step one: enter through the balloon corridor.” Molly: “Those balloons are booby-trapped with helium and optimism.” Bill: “Step two: distract the parent with magic.” Molly: “You can’t do magic.” Bill: “I can do disappearing cupcakes. Watch.” Molly: “Cupcakes disappear when you eat them.” Bill: “That’s the magic of hunger.” Molly: “Okay, plan B: we sing the birthday song backwards.” Bill: “Backwards? That’s just a very polite threat.” Molly: “Shh. The gifts are in the playroom.” Bill: “Playroom means toys and witnesses with sticky hands.” Molly: “No touching. We only steal presents, not innocence.” Bill: “Speak for yourself. I once stole a Lego and still feel guilty about the minifigure’s career prospects.” Molly: “Now focus. Grab the wrapped boxes, slide them into our sack, and leave the unwrapped trash behind.” Bill: “What unwrapped trash?” Molly: “The educational confidence of the parents.” Bill: “Got it. If anyone asks, we’re ‘Gift Consultants.’” Molly: “Perfect. Consultants never steal; they merely relocate joy.” Bill: “Then let’s steal all the joy, starting with the biggest train.” Molly: “On my count: three…” Bill: “Watch the frosting cannons; I’ll handle the gift wrap, Molly, quietly.” Bill: “Two…” Molly: “One—”
created_at
"2026-03-18T10:54:57.223Z"
3/18/2026, 10:54:57 AM
model
"openai/gpt-5.4-nano-20260317"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
12864
moderation_latency
(null)
generation_time
13183
tokens_prompt
77
tokens_completion
1039
native_tokens_prompt
66
native_tokens_completion
2186
native_tokens_completion_images
0
native_tokens_reasoning
1807
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
0.0027457
router
(null)
provider_responses
0
endpoint_id
"0c835f2e-c18d-4e8c-b245-e1e3bd08b97f"
id
"resp_09e92ea84b946a680169ba848146c08193a770c297fbf0b8ad"
is_byok
false
latency
179
model_permaslug
"openai/gpt-5.4-nano-20260317"
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-1773831297-Hjsx4QhlFWKNMfdI2zje"
upstream_id
"resp_09e92ea84b946a680169ba848146c08193a770c297fbf0b8ad"
total_cost
0.0027457
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
Evaluation details
Result
Evaluator
Details
Meta Data
3.62%
Matches word count
n/a
neededClean
false
words
224
99.94%
Dialogue to Total Word Ratio
Ratio: 88.44%, Deviation: 1.56%
neededClean
false
wordsTotal
225
wordsDialogue
199
51.7824%