NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-02-01M1HHG738HH1TBZEAQKNJ763C
Run Details
Model:
Gemini 3.8 Flash (Reasoning, Medium)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
18.0s
Cost:
$0.021013
Provider:
Google AI Studio
Evaluation summary
Score
Evaluator
Details Link
99.74%
Matches word count
(details)
46.51%
Dialogue to Total Word Ratio
(details)
73.1250%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 70% 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
"Tell me again why we are hitting a bouncy castle," Bill muttered, staring down at the crude blueprint scribbled in blue wax crayon on a greasy diner napkin. "It is not just a bouncy castle, Bill," Molly said, jabbing the napkin with a polished fingernail. "It is an impenetrable fortress of suburban excess. Look at the intel. The target is turning seven. His parents manage a venture capital firm. We are talking high-end drones, imported robotic kits, and sealed cards from guilt-ridden grandparents stuffed with untraceable hundred-dollar bills." "They are children, Molly," Bill said, shaking his head in disbelief. "We used to crack bank vaults in Zurich. Now we are lifting plastic trucks from second-graders." "Zurich had laser grids and armed guards," Molly replied, leaning forward across the booth. "Second-graders have sticky fingers and zero perimeter defense. It is the cleanest score of our professional careers." "And our tactical insertion point? Are we squeezing through the pet door?" "Negative. We walk straight through the front gate disguised as the hired talent." Molly reached beneath the plastic table and hoisted two massive duffel bags. "I secured two high-end entertainer uniforms from a theatrical surplus warehouse." Bill unzipped the nearest canvas bag, his jaw dropping in absolute horror. He stared at the grotesque synthetic fabric, questioning every life choice that had brought him to this booth. "Are these clown suits?" "They are tactical disguise garments." "Molly, this is a neon yellow jumpsuit with giant red polka dots and a ruffled lace collar." "You will wear the polka dots, Bill, or you will forfeit your cut of the bounty," Molly warned, narrowing her eyes. "Now pay attention to the timeline. At fourteen hundred hours, the sheet cake is served. Sugar saturation reaches critical mass. The parents will be barricaded inside the kitchen drinking pinot grigio to drown out the shrieking. That is our breach window." "What about the guard dog?" Bill asked. "A golden retriever named Barnaby," she said smoothly, crossing her arms with smug satisfaction. "I have three thick-cut bacon strips stashed inside my oversized left shoe. The canine will be neutralized before we hit the back patio." "And the loot extraction?" "The presents are stacked on the sunroom buffet," Molly whispered. "While they smack the pinata, we bag the haul and hop the cedar fence into the alleyway." "What if an ankle-biter spots us?" "Deploy the countermeasure," she instructed. "Which is?" "Twist a balloon dachshund so violently that it pops. The acoustic shockwave will cause localized hysteria. While they weep, we sweep." Bill rubbed his temples, sighing deeply under the harsh fluorescent diner lighting. "A popping balloon is not a flashbang grenade, Molly." "To a seven-year-old on red dye forty, it is total psychological warfare." "You are a terrifying human being," Bill muttered, pulling a squeaking red ball from the bag. "This rubber nose pinches my septum." "Suck it up," Molly said, adjusting her bright rainbow wig. "Grab the burlap sacks and honk once for go time."
created_at
"2026-09-02T17:08:39.664Z"
9/2/2026, 5:08:39 PM
model
"google/gemini-3.8-flash-20260902"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1343
moderation_latency
(null)
generation_time
17960
tokens_prompt
75
tokens_completion
2563
native_tokens_prompt
67
native_tokens_completion
5590
native_tokens_completion_images
0
native_tokens_reasoning
4909
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
0.02101275
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788368919-wFhLIbzCXyPZErjt9KFm"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788368919-S6s6Rb65mp2tPUA4I9I8"
upstream_id
"F1iYaqXILOf2jrEP3Yip6Q8"
provider_responses
0
endpoint_id
"d376f781-516d-42f9-936d-46f19abca6cc"
id
"F1iYaqXILOf2jrEP3Yip6Q8"
is_byok
false
latency
1343
model_permaslug
"google/gemini-3.8-flash-20260902"
provider_name
"Google AI Studio"
status
200
total_cost
0.02101275
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
99.74%
Matches word count
n/a
neededClean
false
words
496
46.51%
Dialogue to Total Word Ratio
Ratio: 73.91%, Deviation: 3.91%
neededClean
false
wordsTotal
506
wordsDialogue
374
73.1250%