NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-02-01M1HH63EGTZ1XRX9T941E5PSN
Run Details
Model:
Gemini 3.8 Flash (Reasoning, Medium)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
25.8s
Cost:
$0.034910
Provider:
Google AI Studio
Evaluation summary
Score
Evaluator
Details Link
90.48%
Matches word count
(details)
99.72%
Dialogue to Total Word Ratio
(details)
95.1015%
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
The surveillance van smelled faintly of stale onion rings and high-stakes criminal ambition. Molly lowered her military-grade binoculars, peering through the grease-smeared windshield at the primary target. Across the sunny suburban cul-de-sac sat a two-story pastel McMansion currently vibrating from the acoustic terror of twelve hyperactive seven-year-olds. On the front lawn, an enormous lime-green bouncy castle wheezed against its yellow nylon tethers like a dying dragon, surrounded by dozens of discarded juice pouches and glittery plastic tiaras. Beside her on the ripped vinyl bench seat, Bill wrestled clumsily with the elastic strap of an oversized, brightly colored papier-mache donkey head. "Why am I the pinata again? It chafes my sensitive neck." Molly tapped the tactical blueprint spread across the dashboard, which was actually just a greasy pepperoni pizza box covered in frantic red marker arrows. "Because nobody questions a donkey entering a backyard. Focus, Bill. At fourteen-hundred hours, the mother presents the cake. That buys us six minutes while their blood sugar spikes." "Then we breach the sunroom from the windows?" Molly clicked her ballpoint pen with ruthless, professional precision. "Target package alpha. Sixteen wrapped boxes, three gift bags, and that motorized miniature jeep." Bill pushed the donkey snout upward to reveal his hopeful eyes. "Dibs on driving the jeep. It has real rubber tires, Molly." She glared at him with the cold, unyielding fury of a seasoned criminal mastermind. "Nobody drives the jeep, idiot. We bag the haul and scale the fence before the clown finishes his balloon poodle. Do you have perimeter suppression?" Bill reached proudly into the deep pockets of his baggy trench coat, producing a squeaky yellow rubber chicken alongside four foil-wrapped pouches. He brandished them in the dim cabin light like high-grade military ordnance. "Chicken for the golden retriever, apple juice for the kids. Flawless diversion." Molly checked her stopwatch, steeling her nerves against the terrifying prospect of sticky fingers. "If any toddler grabs your leg, play dead. Understood?" Bill pulled the papier-mache head back down over his ears. "Understood. Can I hit the bouncy castle on the way out?" Her icy glare could easily have melted titanium. "Touch that vinyl and I leave you to the PTA." A muffled sigh echoed from inside the hollow cardboard skull. "You take all the joy out of crime, Molly." Molly gripped the door handle with grim determination, her knuckles whitening. "Masks on. Move out." She pulled down her black knit ski mask and kicked the rusted side door open. Outside, the blinding midday sun beamed down on the chaotic battleground of spilled fruit punch, rogue streamers, and inflatable obstacle courses. A small fortune in unopened construction sets and remote-controlled drones waited completely unguarded inside the climate-controlled sunroom, protected by nothing more formidable than a sleepy golden retriever and a visibly hungover balloon clown. Molly slipped swiftly into the neighbor's dense hydrangea bushes, fully aware that history always favored the bold, even when accompanied by an armed donkey.
created_at
"2026-09-02T17:03:08.248Z"
9/2/2026, 5:03:08 PM
model
"google/gemini-3.8-flash-20260902"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1406
moderation_latency
(null)
generation_time
25801
tokens_prompt
75
tokens_completion
3254
native_tokens_prompt
67
native_tokens_completion
9296
native_tokens_completion_images
0
native_tokens_reasoning
8633
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
0.03491025
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788368588-WKkC7LFD4aErRVSXXM0A"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788368588-yex889VslacMzzZWDHcU"
upstream_id
"zFaYaoX3EtnC39IPyuPvgQE"
provider_responses
0
endpoint_id
"d376f781-516d-42f9-936d-46f19abca6cc"
id
"zFaYaoX3EtnC39IPyuPvgQE"
is_byok
false
latency
1388
model_permaslug
"google/gemini-3.8-flash-20260902"
provider_name
"Google AI Studio"
status
200
total_cost
0.03491025
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Google AI Studio"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
90.48%
Matches word count
n/a
neededClean
false
words
490
99.72%
Dialogue to Total Word Ratio
Ratio: 30.24%, Deviation: 0.24%
neededClean
false
wordsTotal
506
wordsDialogue
153
95.1015%