NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-20-01KHXPXEAJNRCNSQEMHE6R30FX
Run Details
Model:
Mistral Large 3
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
13.8s
Cost:
$0.001118
Provider:
Mistral
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
31.33%
Dialogue to Total Word Ratio
(details)
15.6645%
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 back room of the Tipsy Turtle pub smelled like stale beer and regret, which was fitting, since that’s exactly what Molly and Bill were serving up tonight. Molly leaned against the sticky table, spinning a butter knife between her fingers like it was a lockpick. Bill, meanwhile, was hunched over a crumpled napkin, scribbling what he swore was a "foolproof plan" but looked more like a toddler’s finger-painting of a Rube Goldberg machine. "You’re telling me," Molly said, flicking the knife so it stuck in the table an inch from Bill’s hand, "that our grand scheme is to waltz into a six-year-old’s birthday party and just… take the presents?" Bill didn’t flinch. "Not *just* take them. We’re gonna *blend*." Molly squinted. "Blend." "Yeah. Like a smoothie. But for crime." She pinched the bridge of her nose. "Bill, smoothies don’t commit felonies." "Exactly! That’s why it’ll work. No one suspects the smoothie." Molly sighed and plucked the knife from the table, twirling it again. "Walk me through this. Step by step. And try not to say ‘smoothie’ again." Bill cleared his throat and smoothed out the napkin. "Alright. Step one: we get jobs at the party." Molly blinked. "Jobs." "Yeah. Like, you’re the magician. I’m the… uh… balloon animal guy." "The *what* now?" "You know, those long balloons you twist into dogs and swords and stuff." Molly stared at him. "Bill, I once saw you try to tie your own shoelaces and it ended with you sitting on the floor for ten minutes. You think you can make a balloon *swan*?" Bill waved a hand. "Pfft, how hard can it be? You just twist and—bam!—art." Molly groaned. "Fine. So we’re a magician and a balloon artist. Then what?" "Then we mingle. We do our thing. We make the kids happy. And while everyone’s distracted by the *amazing* balloon giraffe I’m gonna make—" "You don’t even know what a giraffe *looks* like." "—we slip into the present room, grab the loot, and stroll out like we own the place." Molly tapped the knife against her palm. "And the parents? The security? The *clown*?" Bill grinned. "That’s the beauty of it. No one suspects the entertainment. We’re *invisible*." Molly considered this. Then she considered the last time Bill had tried to be "invisible" and ended up setting a curtain on fire. "Alright. But we’re doing this my way." Bill perked up. "Oh yeah? What’s your way?" Molly leaned in, her voice dropping to a conspiratorial whisper. "We bring a *real* magician. Someone who can actually make things disappear." Bill’s face fell. "But that’s *cheating*." Molly smirked. "No, Bill. That’s *professionalism*." Bill opened his mouth to argue, then closed it. "…Fine. But I’m still doing the balloon animals." Molly sighed. "God help us all."
created_at
"2026-02-20T14:22:59.934Z"
2/20/2026, 2:22:59 PM
model
"mistralai/mistral-large-2512"
1/1/2512, 12:00:00 AM
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
343
moderation_latency
(null)
generation_time
13750
tokens_prompt
77
tokens_completion
753
native_tokens_prompt
69
native_tokens_completion
722
native_tokens_completion_images
(null)
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)
origin
"https://ncbench.com/"
is_byok
false
finish_reason
"stop"
native_finish_reason
"stop"
usage
0.0011175
router
(null)
provider_responses
0
endpoint_id
"fa718841-07ea-4b7e-b8a3-6069545c7e6a"
id
"e09eda872d584e6ba8ed2883e89f0f54"
is_byok
false
latency
342
model_permaslug
"mistralai/mistral-large-2512"
1/1/2512, 12:00:00 AM
provider_name
"Mistral"
status
200
api_type
"completions"
id
"gen-1771597379-8gwLObGKIIINUGgHg2CA"
upstream_id
"e09eda872d584e6ba8ed2883e89f0f54"
total_cost
0.0011175
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Mistral"
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
467
31.33%
Dialogue to Total Word Ratio
Ratio: 54.82%, Deviation: 4.82%
neededClean
false
wordsTotal
467
wordsDialogue
256
15.6645%