NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYCX79B98F8FATFNW0KYE3T0
Run Details
Model:
Gemini 3.6 Flash (Reasoning, Minimal)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
5.3s
Cost:
$0.004938
Provider:
Google AI Studio
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.06%
Dialogue to Total Word Ratio
(details)
0.0320%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 50% 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 party was happening at three o'clock, which meant Molly and Bill had precisely thirty minutes to review the blueprint of a bouncing castle. They sat in the back of a rented van that smelled strongly of damp dog and pine needles. Molly adjusted her tactical balaclava, which was bright pink and featured bunny ears. Bill looked at her with deep concern. "Are you sure about the headwear? It lacks subtlety." "It creates misdirection, Bill. The kids will think I am the entertainment, right up until I swap the birthday boy's Nintendo with a spray-painted brick." Molly slapped the blueprint onto the dashboard. "Now, focus. What is the entry point?" "The slide," Bill said, pointing at the diagram with a half-eaten pretzel. "If we crawl up the yellow chute, we bypass the mothers drinking prosecco by the gazebo." "Good. What about the guard dog?" "It’s a twelve-year-old Pug named Mr. Sprinkles. I have prepared a bribe." Bill pulled a greasy brown paper bag from his coat and produced a single, cold cocktail sausage. "He will be neutralized instantly." Molly nodded approvingly. "Ruthless. I like it. Now, let’s talk about the primary target. Where is the high-value stash?" "According to my reconnaissance, the pile is situated on the trestle table beside the musical chairs arena. We are looking at six Lego sets, three remote-controlled helicopters, and an envelope containing forty pounds from a distant aunt." "Jackpot," Molly whispered, her eyes widening. "What is the exit strategy if the bouncy castle loses pressure?" "We roll," Bill said gravely. "We roll to the hedge line and deploy the smoke bombs." "Are those actual smoke bombs?" "They are glitter cannon party poppers, but at close range, the sparkle is blinding." Molly sighed and leaned back against the headrest. "Bill, this is our biggest score yet. No room for error. If a seven-year-old spots us, what do you do?" "I drop to the floor and pretend to be a oversized stuffed bear until they lose interest." "Correct. And if the birthday boy demands his presents back?" "I tell him the capitalism monster took them and offer him a fruit shoot." "Flawless," Molly declared, checking her watch. "It’s time. Put on your clown nose." Bill pulled a red foam ball from his pocket and pressed it onto his nose with a wet squelch. "I hate this part of the job." "You love it, and you know it. Now grab the burlap sacks." "They say 'Santa's Magic Mail' on the side, Molly." "It’s psychological warfare, Bill. They won't suspect a thing." Molly slammed the van door open, ducking low as the sound of screaming children and pop music drifted over the wooden fence. They crept toward the perimeter, ready to pull off the greatest playground heist of the century.
created_at
"2026-07-25T15:10:12.337Z"
7/25/2026, 3:10:12 PM
model
"google/gemini-3.6-flash-20260721"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
468
moderation_latency
(null)
generation_time
5245
tokens_prompt
75
tokens_completion
701
native_tokens_prompt
67
native_tokens_completion
645
native_tokens_completion_images
0
native_tokens_reasoning
0
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
0.004938
router
(null)
provider_responses
0
endpoint_id
"5d6d133d-a953-4781-a835-7ee76f4e1388"
id
"1NFkarTLGOaV_uMPiqjl-Aw"
is_byok
false
latency
468
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-1784992212-3a0eILBey1pga43VK14v"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1784992212-4vxshkJLSI9abCCDH61e"
upstream_id
"1NFkarTLGOaV_uMPiqjl-Aw"
total_cost
0.004938
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
0.00%
Matches word count
n/a
neededClean
false
words
462
0.06%
Dialogue to Total Word Ratio
Ratio: 62.13%, Deviation: 12.13%
neededClean
false
wordsTotal
470
wordsDialogue
292
0.0320%