NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-10-07-01M4AT5NN8M99KKPEPH8TT5RC4
Run Details
Model:
Mistral Large 4.0 (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
5m 38s
Cost:
$0.042293
Provider:
Mistral
Evaluation summary
Score
Evaluator
Details Link
3.62%
Matches word count
(details)
69.39%
Dialogue to Total Word Ratio
(details)
36.5088%
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
Molly spread the crayon-drawn map across the kitchen table, tracing the line labeled Main Entrance. Bill leaned in, squinting at the glitter stains obscuring the street names, his nose nearly touching the paper. "This is the Henderson residence," he said, tapping the paper. "The target acquired their shipment of loot three days ago. High-end toys, plushies, and a gift-wrapped drone the size of a small dog." Molly nodded, adjusting her tactical fanny pack, her eyes narrowing at the impossible timeline. "We hit them at 2:15 PM, right after the cake. The sugar high will distract the parents, masking our escape." Bill scratched his chin, a habit from watching too many noir films. "You sure about the bounce house? Last time we used an inflatable, we got stuck when a toddler sat on the exit." "That was a unicorn," Molly said. "This is a medieval castle. Structurally different." He pulled out a clipboard, the metal clip rusted from a previous job involving mayonnaise. "Okay, phase one: infiltration. I've acquired party hats. Genuine Mylar, not cheap latex. We blend in with the kids." Molly crossed her arms, her posture suggesting she was already regretting this enterprise. "Bill, we're forty-three. We have gray hair. We cannot blend in with kids." "We have the hats," he insisted. "I practiced the voice. Listen." He cleared his throat in a high, squeaky voice. "Timmy wants more juice, Mommy!" Molly winced, visions of lawsuits dancing in her head. "That sounds like a dying seagull." "It sounds festive," Bill said. "Phase two: the grab. We wait until the piñata. When candy rains down, the children will frenzy. Chaos is our ally." "Or a tactical nightmare," Molly said. "Have you seen what happens when you deny a child a tiara? It's like Lord of the Flies with frosting." "We go in hard and fast," Bill said. "Two minutes, tops. Grab the presents, avoid the piñata stick, and vanish before anyone notices." "What about the goody bags?" Molly asked. "They're already distributed." "Already handled," Bill said. "I befriended the caterer. He thinks we're magicians. We arrive, grab the presents, and exit stage left." "Bill, we're thieves," Molly said. "Not children's entertainers." "We're professionals," Bill corrected. "There's a difference. Professionals adapt. Now, the getaway vehicle. I've reserved an ice cream truck from Fifth Street." Molly stared at him, wondering if he'd hit his head during the drive over. "A what? You reserved a what?" "An ice cream truck," Bill said proudly. "It's inconspicuous. Parents think treat, not felony. We can use the freezer for loot and hide the drone in the soft-serve." Molly paced the kitchen, kicking a stray Lego brick under the table. "You want to steal presents from kids in an ice cream truck while playing that jingle." "I want to steal presents from kids in an ice cream truck while playing the jingle," Bill said. "It's called ambiance. It soothes the victims." Molly sighed, accepting her fate as an accomplice to childhood larceny. "Fine. If one of those kids recognizes us, I'm blaming your voice." "The voice is an asset," Bill said. "It's disarming." "It's terrifying. But okay. Let's steal some Legos."
created_at
"2026-10-07T09:12:35.761Z"
10/7/2026, 9:12:35 AM
model
"mistralai/mistral-large-4-0-20261006"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2274
moderation_latency
(null)
generation_time
338387
tokens_prompt
75
tokens_completion
17087
native_tokens_prompt
67
native_tokens_completion
20214
native_tokens_completion_images
(null)
native_tokens_reasoning
16274
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
0.04229282
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1791364355-gkwjcqds5NIBnrH2DQ2M"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1791364355-oTRQ976hhZWVCmMV61uj"
upstream_id
"8d072a75fc914d2fa1f40c24669522ad"
provider_responses
0
endpoint_id
"19b3b5d7-23f6-4b63-8109-64a6cc37d996"
id
"8d072a75fc914d2fa1f40c24669522ad"
is_byok
false
latency
564
model_permaslug
"mistralai/mistral-large-4-0-20261006"
provider_name
"Mistral"
status
200
total_cost
0.04229282
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Mistral"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
3.62%
Matches word count
n/a
neededClean
false
words
524
69.39%
Dialogue to Total Word Ratio
Ratio: 67.30%, Deviation: 2.70%
neededClean
false
wordsTotal
529
wordsDialogue
356
36.5088%