NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-20-01KHYB2W3H1MJPRFZ0ZD87WVWA
Run Details
Model:
Ministral 3 8B
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
3.3s
Cost:
$0.000101
Provider:
Mistral
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
54.49%
Dialogue to Total Word Ratio
(details)
27.2435%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 70% 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 neon sign outside the *Whimsy Wonderland* toy store flickered like a drunk firefly, casting a sickly glow over the parking lot. Molly adjusted her sunglasses—even though it was dusk—and leaned against the hood of Bill’s beat-up van, which smelled suspiciously like last night’s takeout and regret. Bill kicked a pebble down the street, then sighed. "Okay, so the plan is still to break into the *kids’ birthday party* at the community center?" Molly crossed her arms. "Yeah, Bill. The plan is *still* to steal all the presents before the parents even realize they’re missing. Which, by the way, is why we’re not calling it a ‘heist.’ We’re calling it a *gift exchange.*" Bill groaned. "A gift exchange where we take *everything* and leave them with a single deflated balloon and a note that says *‘Happy Birthday, you broke people.’*" "Exactly." Molly grinned. "And we’ll be long gone by the time the first kid starts crying over their missing LEGO set." Bill rubbed his temples. "What’s the security like? I saw a security guard outside the toy store earlier." Molly waved a hand. "Pfft. He’s the kind of guy who’d fall asleep if you handed him a cupcake. Besides, we’re not breaking into the toy store. We’re breaking into the *party.*" Bill’s eyes widened. "Ohhh, you mean the *actual* party? With the *actual* kids?" "Yep." Molly popped the P. "And here’s the thing—kids are *terrible* at security. They’ll either be too busy opening presents to notice us, or they’ll think we’re part of the entertainment." Bill shook his head. "You’re insane." "Brilliant," Molly corrected. "So here’s how it goes: I’ll distract the parents with a fake ‘lost puppy’ story while you slip in through the side door. Once inside, you grab all the presents—*everything*—and we meet back at the van. Then we split the haul before the cops even show up." Bill frowned. "What if someone *sees* us?" Molly smirked. "Then we tell them we’re the new clowns. Kids love clowns. Parents *hate* clowns. Perfect cover." Bill stared at her. "You’re *not* serious about the clown thing." "Wouldn’t you rather be arrested for *clown-related crimes* than *grand theft present*?" Bill sighed, defeated. "Fine. But if we get caught, I’m blaming you." Molly grinned. "Deal. Now, you ready to make some *very* happy kids *very* unhappy?" Bill adjusted his gloves. "Let’s do this."
created_at
"2026-02-20T20:15:29.402Z"
2/20/2026, 8:15:29 PM
model
"mistralai/ministral-8b-2512"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
209
moderation_latency
(null)
generation_time
2940
tokens_prompt
77
tokens_completion
644
native_tokens_prompt
69
native_tokens_completion
605
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
0.0001011
router
(null)
provider_responses
0
endpoint_id
"c85d26e1-1e0c-449b-9775-2afce7ae510b"
id
"0bfbb796036348408dd4a619b6955fbb"
is_byok
false
latency
209
model_permaslug
"mistralai/ministral-8b-2512"
provider_name
"Mistral"
status
200
api_type
"completions"
id
"gen-1771618529-9NGZ78rXHTSBGXgPGmDo"
upstream_id
"0bfbb796036348408dd4a619b6955fbb"
total_cost
0.0001011
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
396
54.49%
Dialogue to Total Word Ratio
Ratio: 73.48%, Deviation: 3.48%
neededClean
false
wordsTotal
396
wordsDialogue
291
27.2435%