NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-05-10-01KR91G8JXZ0PXZ101ZHY4KJ3V
Run Details
Model:
Gemma 4 31B (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
13m 4s
Cost:
$0.003315
Provider:
Parasail
Evaluation summary
Score
Evaluator
Details Link
60.28%
Matches word count
(details)
15.85%
Dialogue to Total Word Ratio
(details)
38.0637%
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 spread the blueprints of the suburban living room across the grease-stained folding table, her eyes gleaming with professional intensity. "The target is a seven-year-old's birthday bash," she whispered. Bill stared at the sprawling blueprints of the suburban residence, looking profoundly confused as he chewed on a splintering toothpick. "Why are we hitting a kid's party?" "High volume, low security," Molly snapped, tapping a silver pen aggressively against the map, nearly poking a hole through the dining room layout. "The presents are concentrated in one area. We strike during the cake ceremony." Bill scratched his chin with a calloused hand, thinking deeply about the logistics of the covert operation. "What about the parents?" "Distracted by the chaos. Parents at these things are basically zombies," she explained with a smug smirk that suggested she had already won. "Once the candles are lit, every adult in the room will be staring at the cake like it's a holy relic." "And the kids?" Bill asked, his voice sounding slightly concerned as he looked toward the window. "They're fast. Like, scary fast." Molly pointed to a large red circle on the map with absolute, unwavering certainty. "That is the bounce house. We lure them in with a handful of glitter and a loud noise. While they're bouncing, we clear the gift table." Bill frowned, leaning in to inspect the perimeter for any hidden security cameras. "Do I need a disguise?" "Obviously. You're going in as a clown," Molly said, sliding a neon-orange wig across the table. Bill looked horrified, recoiling from the synthetic hair as if it were a venomous snake. "I can't do clowns, Molly. I have a professional reputation." "Your reputation is being the guy who once tripped over his own shoelaces during a jewelry heist," she retorted. "Put on the wig." "What about you?" he asked, eyeing her sharp professional attire and the way she carried herself. "I'll be the entertainment coordinator," she replied. "I have the clipboard. No one questions a woman with a clipboard." Bill sighed, tentatively touching the wig with one finger. "Is there a backup plan if the magician shows up?" "We bribe him with a pack of cigarettes and a fake endorsement," Molly decided. "He's a hack; he'll fold instantly." "What if there are dogs?" Bill wondered, his voice trembling. Molly rolled her eyes with a heavy, theatrical sigh. "It's a Golden Retriever, Bill. It'll probably help us carry the bags." "I just don't like the risk," he murmured, imagining a canine onslaught. "The risk is minimal," she countered. "The reward is three Lego sets, a miniature drone, and possibly a gift card to Toys R Us." Bill perked up, his eyes widening. "Wait, really?" "According to the invitation, the theme is 'High Tech Adventure,'" she noted. Bill grabbed the wig with sudden, greedy enthusiasm. "Fine. But I'm not making any balloon animals." "Just grab the boxes and run," Molly commanded. "We exit through the sliding glass door before the piƱata breaks." "What if I get stuck in the bounce house?" Bill asked. "Then you're the new birthday boy," Molly sighed.
created_at
"2026-05-10T13:33:52.101Z"
5/10/2026, 1:33:52 PM
model
"google/gemma-4-31b-it-20260402"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
704
moderation_latency
(null)
generation_time
783704
tokens_prompt
77
tokens_completion
7753
native_tokens_prompt
82
native_tokens_completion
8259
native_tokens_completion_images
(null)
native_tokens_reasoning
6932
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
(null)
usage
0.00331508
router
(null)
provider_responses
0
endpoint_id
"9d4644e9-a5e5-41f4-b810-d75f484d2c24"
id
"chatcmpl-ac2c50b1f9dcd7d6"
is_byok
false
latency
415
model_permaslug
"google/gemma-4-31b-it-20260402"
provider_name
"Parasail"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.1; linux; x64))"
http_referer
(null)
request_id
"req-1778420032-FwIdleDISdOXKMACNmun"
session_id
(null)
api_type
"completions"
id
"gen-1778420032-RnNzgCOqqLR0ia18RYJc"
upstream_id
"chatcmpl-ac2c50b1f9dcd7d6"
total_cost
0.00331508
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Parasail"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
60.28%
Matches word count
n/a
neededClean
false
words
515
15.85%
Dialogue to Total Word Ratio
Ratio: 56.07%, Deviation: 6.07%
neededClean
false
wordsTotal
519
wordsDialogue
291
38.0637%