NC Bench
Overview
About
Tests
Models
Vendors
Compare
Run: 2026-02-25-01KJ95N7229EWZGSSBQ65GBQW2
Run Details
Model:
Gemini 3 Flash (Preview, Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
17.2s
Cost:
$0.008604
Provider:
Google AI Studio
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
0.0000%
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
Molly leaned over the dashboard, her binoculars pressed so hard against her eyes she was going to have rings for a week. Outside the tinted windows of the nondescript transit van, the target loomed: Number forty-two, Wisteria Lane. The balloons tied to the mailbox whipped violently in the breeze, like colorful distress signals. They’ve upgraded to biometric security, Molly whispered, her voice tight with professional tension. It’s a dog, Bill said, his voice muffled by a mouthful of ham and rye. It’s just a golden retriever, Molly. It sleeps in a wicker basket and farts in its sleep. I saw it during the pre-op surveillance. That beast is a Furry Motion Sensor with Teeth, she snapped, grabbing a tablet and swiping through blueprints of the garden. Did you pack the distraction? I’ve got the bag of premium beef jerky right here, Bill said, patting a tactical pouch on his thigh. Guaranteed to disable any canine within a fifty-yard radius. Good. Now, look at the objective. She pointed a manicured finger at the gift table visible through the hedge. It was a mountain of neon-wrapped boxes, topped with a ribbon so large it likely had its own zip code. In the center sat the heavy-duty prize: the Ultimate Galactic Fortress Lego Set. It’s beautiful, Bill breathed, a spec of mustard landing on his black turtleneck. I mean, the resale value is astronomical on the secondary brick market. Focus, Bill. We move in at fourteen-hundred hours, right after the cake is served. Why then? Bill asked. Why not during the puppet show? Sugar crashes, you idiot. The primary targets will be lethargic, crying, or experiencing total sensory overload. The parents will be barricaded in the kitchen behind a wall of Chardonnay and regret. It’s the perfect window for a clean extraction. What about the clown? Bill squinted at a man currently failing to make a balloon giraffe. He looks shifty. Peppino the Perilous? Molly pulled a dossier from her bag. Real name: Gary. He has a trick knee and a crippling fear of humidity. If he tries to intervene, remind him that his balloon animals are structurally unsound and his makeup is smearing. I don’t like his shoes, Bill muttered. You can hide a lot of weaponry in a size twenty-four boot. It’s tactically unsound. Just stick to the plan. You enter through the bouncy castle. Why the bouncy castle? That is a high-visibility zone! It’s the blind spot! Molly hissed. The mesh siding provides perfect visual concealment from the patio. You bounce, you grab at the apex of your jump, and you bounce back out. I get motion sickness on those things, Molly. Last time I was in an inflatable structure, I threw up on a bouncy dragon. Then take a Dramamine and man up. This is the big one, Bill. We’re talking three Power Wheels and a motorized Barbie Jeep with working headlights. How are we going to fit a Jeep in the van? We’ll tow it, Bill. We’ll tow it behind us like a trophy of war. Bill sighed, checking his tactical watch. It had a picture of a smiling cartoon sun on the face. Ten minutes until the piñata ceremony begins. Remember, once the stick hits the donkey, it’s every thief for themselves. What if there’s a magician? Bill asked, his voice trembling. You didn’t mention the possibility of a magician. If a man in a cape tries to make our exit vanish, you hit him with the tactical juice boxes. Copy that. Juice boxes primed and ready for deployment. Molly pulled her aviators down, the plastic clicking into place. Let’s go make some six-year-olds very confused.
created_at
"2026-02-25T01:12:17.735Z"
2/25/2026, 1:12:17 AM
model
"google/gemini-3-flash-preview-20251217"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1768
moderation_latency
(null)
generation_time
17160
tokens_prompt
77
tokens_completion
1836
native_tokens_prompt
67
native_tokens_completion
2857
native_tokens_completion_images
0
native_tokens_reasoning
2036
native_tokens_cached
0
num_media_prompt
(null)
num_input_audio_prompt
(null)
num_media_completion
0
num_search_results
(null)
origin
"https://ncbench.com/"
is_byok
false
finish_reason
"stop"
native_finish_reason
"STOP"
usage
0.0086045
router
(null)
provider_responses
0
endpoint_id
"6e82f652-2693-4b60-bbe0-7362d6e345f0"
is_byok
false
latency
1768
model_permaslug
"google/gemini-3-flash-preview-20251217"
provider_name
"Google AI Studio"
status
200
api_type
"completions"
id
"gen-1771981937-H674wlmsDl1qoRX7uY7O"
upstream_id
(null)
total_cost
0.0086045
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Google AI Studio"
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
609
0.00%
Dialogue to Total Word Ratio
Ratio: 0.00%, Deviation: 50.00%
neededClean
false
wordsTotal
618
wordsDialogue
0
0.0000%