NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-11-19-01M7SPY0JMXFP1W60SSQ553WWJ
Run Details
Model:
Gemini 4 Neon (Reasoning, Medium)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
1m 9s
Cost:
$0.124551
Provider:
Google
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
98.70%
Dialogue to Total Word Ratio
(details)
99.3494%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 30% 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 tapped a sharpened pencil against the laminated aerial photograph of 42 Maple Street, her brow furrowed with the intensity of a general plotting the invasion of Normandy. Across the battered kitchen table, Bill stared nervously at a plastic green army man representing a seven-year-old boy named Alistair. The table was buried under surveillance photos, a stolen invitation embossed with cartoon trucks, and three half-eaten glazed donuts. It was the biggest score of their dismal careers, provided they could survive the sugar-fueled chaos of a suburban Saturday afternoon. "Alistair's father invented an algorithm that predicts stock market crashes," Molly said, sliding a shiny catalog across the linoleum. "His mother owns three vineyards. These are not ordinary birthday gifts, Bill. We are talking about vintage robots, miniature electric sports cars, and an unboxing table worth forty grand." Bill swallowed hard and picked up a grainy photograph taken from the bushes outside the Henderson estate. It showed a sprawling green lawn dominated by a massive inflatable castle shaped like a medieval dragon. Near the patio, a long folding table groaned under a mountain of brightly wrapped boxes topped with enormous satin bows. To Bill, the gleaming paper looked less like an easy payday and more like a bear trap waiting to snap shut. "What about the security detail?" Bill asked, pointing a trembling finger at a separate photo of a man in oversized polka-dot trousers. "The intel says they hired Bubbles the Magnificent. Clowns have erratic patrol patterns, Molly. You cannot predict a man who pulls endless scarves out of his throat." Molly rolled her eyes and nudged a salt shaker toward the inflatable dragon. She had already spent six hours memorizing the itinerary she had fished out of the Hendersons' recycling bin. Every minute was accounted for, from the arrival of the petting zoo to the inevitable meltdown over who got the corner piece of frosting. She leaned forward, whispering conspiratorially. "Bubbles is booked for a thirty-minute balloon-animal set in the front yard at two o'clock," Molly explained. "That draws the parents and the kids away from the patio. You slip over the back hedge in the catering uniform, load the boxes into the laundry van, and vanish before they even blindfold the birthday boy." Bill frowned, tapping his chin as he mentally walked through the extraction route. The plan was deceptively simple, which in his experience meant at least three things would catch fire before they reached the getaway vehicle. Still, the rent was due on Tuesday, and his landlord was far more terrifying than a mob of sugared-up first graders. He looked back at the map, tracing the path from the gift table to the dessert buffet with sudden, intense curiosity. "One last question," Bill muttered, reaching for a donut. "If the coast is clear, do I have time to grab a slice of cake?" Molly sighed and pinched the bridge of her nose, wondering why she never worked solo. "Only if it is chocolate, Bill. Grab two plates."
created_at
"2026-11-19T10:51:41.019Z"
11/19/2026, 10:51:41 AM
model
"google/gemini-4-neon-20261118"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1854
moderation_latency
(null)
generation_time
68672
tokens_prompt
75
tokens_completion
2482
native_tokens_prompt
97
native_tokens_completion
8284
native_tokens_completion_images
0
native_tokens_reasoning
7661
native_tokens_cached
0
num_media_prompt
(null)
num_input_audio_prompt
(null)
num_media_completion
0
num_search_results
(null)
num_fetches
(null)
web_search_engine
(null)
origin
"https://ncbench.com/"
is_byok
false
finish_reason
"stop"
native_finish_reason
"STOP"
service_tier
"default"
usage
0.124551
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.22.2; linux; x64))"
http_referer
(null)
request_id
"req-1795085501-n15N6ehfi9OI79m8WS46"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1795085501-rZusB40EAzpyqV2XF206"
upstream_id
"vdT-auKtBf3J4_UPpPfWGQ"
provider_responses
0
endpoint_id
"cbc1d01f-e431-41ec-9abb-36c1a5b10cb4"
id
"vdT-auKtBf3J4_UPpPfWGQ"
is_byok
false
latency
1854
model_permaslug
"google/gemini-4-neon-20261118"
provider_name
"Google"
status
200
total_cost
0.124551
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Google"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
100.00%
Matches word count
n/a
neededClean
false
words
500
98.70%
Dialogue to Total Word Ratio
Ratio: 30.51%, Deviation: 0.51%
neededClean
false
wordsTotal
508
wordsDialogue
155
99.3494%