NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYCTMWKE6GJV88513KGTT8G8
Run Details
Model:
Gemini 3.6 Flash (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
44.4s
Cost:
$0.077283
Provider:
Google AI Studio
Evaluation summary
Score
Evaluator
Details Link
99.92%
Matches word count
(details)
98.61%
Dialogue to Total Word Ratio
(details)
99.2664%
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 blueprint laid across the grease-stained kitchen table was drawn in pastel crayons, depicting a suburban backyard with terrifying strategic detail. Molly tapped a glittery pink marker against the paper, her eyes narrowing beneath tactical goggles that were entirely unnecessary for indoor planning. Beside her, Bill adjusted his oversized ski mask, which kept catching on his nose ring. The target was high-risk and high-reward: seven-year-old Leo’s birthday extravaganza. The haul was estimated to be worth thousands in untraceable LEGO sets, remote-controlled monster trucks, and at least three wrapped boxes that felt suspiciously like cold, hard cash from overly generous grandparents. "We breach through the bouncy castle at fourteen hundred hours sharp," Molly whispered, tapping the crude drawing. "The sugar rush will be reaching its absolute peak then." "I still think this plan is crazy," Bill muttered. "They have juice boxes, Molly." Bill shivered, remembering their last disastrous attempt at a nursery school heist, where a rogue toddler locked him inside a plastic playhouse for two miserable hours. Molly slammed her fists onto the table, startling him into silence. "Focus, Bill! Capri Suns are tactical hazards, but we carry straw-blockers. The real obstacle is the clown." "The clown?" "Twinkles. Ex-circus. He can fold a balloon sword fast enough to take your eye out. Distract him with the pinata stick while I sweep the gift table." "What if the scary mother spots us?" "Deploy noise-makers to confuse her sensory perception." She unzipped a black duffel bag, revealing their gear: two pairs of squeaky shoes, three boxes of glitter bombs, and a stealth sack made from a dark blue trash bag. Bill reached into the duffel bag and pulled out a small plastic kazoo, inspecting it with deep mistrust. He blew a soft, sad note into the cheap toy, shaking his head slowly. "Is this tiny kazoo really going to hold off hyperactive children?" "It is not a kazoo, Bill. It is acoustic warfare," Molly said, attaching her fake mustache. "We drop in, toss loot over the fence, and hit the getaway tricycle before cake." "Who gets the light-up scooter?" "Fifty-fifty split on all the wheeled toys," Molly said coldly. "Now put on your party hat. It is go time." Bill sighed, pulling a pointy cone of cardboard over his ski mask and fastening the elastic band under his chin. It pinched horribly, but field operations required sacrifices. Molly strapped on her glittery tool belt, stuffed three juice pouches into her tactical pockets for emergency rations, and gave the crayon blueprint one final, dramatic salute. Outside, the distant sound of squeal-filled laughter echoed through the quiet suburban cul-de-sac, signaling that the bouncy castle had just been inflated. The defensive perimeter was set, but the unsuspecting guards were entirely oblivious to the impending chaos. Molly clicked her flashlight on and off twice, signaling the start of the drop sequence. They tiptoed toward the back door, synchronized and absurd, ready to pull off the greatest toy heist the neighborhood had ever seen.
created_at
"2026-07-25T14:25:12.325Z"
7/25/2026, 2:25:12 PM
model
"google/gemini-3.6-flash-20260721"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1507
moderation_latency
(null)
generation_time
44358
tokens_prompt
75
tokens_completion
3939
native_tokens_prompt
67
native_tokens_completion
10291
native_tokens_completion_images
0
native_tokens_reasoning
9632
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
0.077283
router
(null)
provider_responses
0
endpoint_id
"5d6d133d-a953-4781-a835-7ee76f4e1388"
id
"SMdkasypGPOB_uMPsbfl6Q4"
is_byok
false
latency
1507
model_permaslug
"google/gemini-3.6-flash-20260721"
provider_name
"Google AI Studio"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.18.0; linux; x64))"
http_referer
(null)
request_id
"req-1784989512-2njOfkvb88sc7NFxBcqH"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1784989512-64dJtyUvWFuuXkiA08Aa"
upstream_id
"SMdkasypGPOB_uMPsbfl6Q4"
total_cost
0.077283
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Google AI Studio"
response_cache_source_id
(null)
data_region
"global"
Evaluation details
Result
Evaluator
Details
Meta Data
99.92%
Matches word count
n/a
neededClean
false
words
497
98.61%
Dialogue to Total Word Ratio
Ratio: 30.53%, Deviation: 0.53%
neededClean
false
wordsTotal
511
wordsDialogue
156
99.2664%