NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1KRRF6F8XHEPD4SQ9N2A3AH
Run Details
Model:
Gemini 3.7 Flash (Reasoning, Medium)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
20.6s
Cost:
$0.021999
Provider:
Google AI Studio
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
30.66%
Dialogue to Total Word Ratio
(details)
15.3311%
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 slammed the tactical blueprint onto the greasy dashboard of their battered surveillance van. The map had been rendered on a Denny’s children's menu in purple crayon, but she treated the document with solemn professionalism. Outside, the target loomed under the brutal glare of the two o’clock sun: a cheerful suburban home flanked by an inflatable bouncy castle shaped like a neon-green brontosaurus. Bill adjusted his heavy binoculars, squinting past the smudged, tinted glass. "Look at the perimeter," Bill muttered. "Three moms armed with juice boxes near the hedge line." "Those are soccer moms, Bill," Molly whispered, leaning over the steering wheel. "Our target is the gift table near the patio. Timmy is turning seven today, which means high-end electronics, remote-controlled drones, and vintage Lego sets." "What about the package in the glitter wrap?" "Leave it. Glitter is a forensic nightmare." Molly smoothed down her disguise, an itchy floral cardigan paired with oversized novelty sunglasses. "We breach during the cake ritual. When the candles are lit, sensory overload hits the room. That is our window of extraction." Bill lowered the binoculars, rubbing his stubbled jaw with genuine concern. "The clown looks twitchy." "He’s handled. Two balloon poodles bought his silence." "And the sticky menace near the sandbox?" "Subject Toddler Alpha. Highly volatile. Avoid contact at all costs." Molly reviewed the inventory stashed between the front seats. Their arsenal consisted of heavy canvas duffel bags, party noisemakers for acoustic disorientation, decoy boxes stuffed with packing peanuts, and a bag of gummy worms meant for emergency toddler bribery. Bill fastened his black tactical vest directly over a bright Hawaiian tourist shirt. "Full tactical, or do we blend?" Bill asked. "We blend. Put on the party hat, Bill." "The elastic band hurts my ears." "Do you want the drone or not?" Molly hissed. Bill sighed heavily, reached into the glove compartment, and retrieved a shiny cone decorated with cartoon dinosaurs. He snapped the elastic cord under his chin with a wince. He looked less like an elite burglar and more like an overgrown convict crashing a preschool graduation. Across the lawn, the sliding patio door swung open. A frazzled father emerged carrying a massive chocolate sheet cake blazing with seven sparkler candles. A chaotic herd of screaming children stampeded toward the picnic benches, leaving the gift table completely unguarded. Molly gripped the van’s door latch, checking her stopwatch. "Operation Juice Box is a go." "God help us if they start musical chairs," Bill said. Molly shoved the door open. They slipped into the humid afternoon air, marching toward their plastic-wrapped fortune with ruthless, joyful precision.
created_at
"2026-09-03T13:53:59.002Z"
9/3/2026, 1:53:59 PM
model
"google/gemini-3.7-flash-20260813"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3993
moderation_latency
(null)
generation_time
20489
tokens_prompt
75
tokens_completion
2407
native_tokens_prompt
67
native_tokens_completion
5853
native_tokens_completion_images
0
native_tokens_reasoning
5270
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
0.021999
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788443639-eYAo2XsQqVxQBuAIezXn"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788443639-Kf7eXsdSAMTBpThcyGMN"
upstream_id
"93uZapKLEYaw1MkP1-f7wAE"
provider_responses
0
endpoint_id
"dd072012-3274-417d-97e3-e290cba37e31"
id
"93uZapKLEYaw1MkP1-f7wAE"
is_byok
false
latency
3993
model_permaslug
"google/gemini-3.7-flash-20260813"
provider_name
"Google AI Studio"
status
200
total_cost
0.021999
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
0.00%
Matches word count
n/a
neededClean
false
words
432
30.66%
Dialogue to Total Word Ratio
Ratio: 34.86%, Deviation: 4.86%
neededClean
false
wordsTotal
436
wordsDialogue
152
15.3311%