NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-10-07-01M4AV8K18YCHAH3W6ZV82ATM8
Run Details
Model:
Mistral Large 4.0 (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
4m 50s
Cost:
$0.037730
Provider:
Mistral
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.30%
Dialogue to Total Word Ratio
(details)
0.1475%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 30% 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 stolen blueprints across the cracked Formica table, her manicured finger tracing the route from the azalea hedge to the sliding glass door of the Henderson residence. Rain lashed against the corrugated metal roof of their hideout, a converted laundromat that smelled perpetually of dryer sheets and desperation. "It's a seven-year-old's birthday party, Bill," she whispered, eyes gleaming. "Thirty kids. Unattended gift table. We walk in wearing party hats, grab the bags, and vanish." Bill adjusted his ski mask, which had been serving as a sleep mask since Tuesday. He examined his reflection in the chrome washing machine, practicing the desperate smile of a man who had lost his dentures. "Party hats?" he asked. "Molly, I'm forty-two with a receding hairline and a knee that pops like a gunshot when I crouch." "Exactly," she grinned. "You'll look like the entertainment." Bill sighed, pulling out his lock picks and examining them like a surgeon checking instruments. The Hendersons had installed a security system last month, but Molly had already disabled the motion sensors using a modified garage door opener and sheer audacity. She had spent three days casing the location, watching from the oak tree across the street, memorizing the patrol patterns of the golden retriever named Biscuit. "What about the grandmother?" Bill asked, pointing at the kitchen. "She has a hip replacement." "She's napping in the recliner," Molly said. "I checked. She snores like a chainsaw." Molly leaned forward until her nose nearly touched the floor plan. She explained that kids' parties were chaos incarnations where parents clustered in the kitchen drinking cheap wine while the gift table sat in the corner, covered in Mylar and hope. They would enter at 2:15, right when the cake arrived, riding the wave of sugar-induced hysteria straight to the presents. "I don't know any magic tricks," Bill muttered, zipping his duffel bag. "I know how to pick locks and regret my life choices." "You can juggle," Molly reminded him. "Sort of. Remember the clown college incident?" "That involved chainsaws," Bill said, not looking up. They fell silent for a moment, listening to the rain. Somewhere in the distance, a siren wailed. Molly reached into her bag and withdrew a glittery unicorn horn headband, placing it on her head with theatrical precision. "Child psychology," she declared. "If a unicorn offers you a juice box, nobody asks questions." "Last time you wore a costume," Bill said, "you got arrested at a Renaissance fair." "I hate children," Bill continued quietly. "I hate their sticky fingers and their ability to sense fear." "Then don't look at them," Molly advised. "Look at the presents. Think about the LEGO sets. The video games." "And the pony," she added. "If we're lucky." Bill adjusted his collar, transforming himself from a thief into the disappointed uncle every child dreaded at family gatherings. "If I have to sing Happy Birthday," he warned, "I'm leaving you in the ball pit." "You won't have to sing," Molly promised. "I'll handle the vocals." "Harmless? I stole a parking meter last week." "Exactly," she said. "You're perfect." Outside, the rain intensified. Molly grabbed her unicorn horn, and Bill checked his lock picks one final time. The heist of the century was about to begin, and neither of them had slept in thirty-six hours.
created_at
"2026-10-07T09:31:40.626Z"
10/7/2026, 9:31:40 AM
model
"mistralai/mistral-large-4-0-20261006"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2472
moderation_latency
(null)
generation_time
289300
tokens_prompt
75
tokens_completion
16224
native_tokens_prompt
67
native_tokens_completion
18031
native_tokens_completion_images
(null)
native_tokens_reasoning
15363
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
0.03773035
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1791365500-O3cYKLQEHjD69g6ls1Yz"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1791365500-LKh2ZJRdwqFsBge8RKJO"
upstream_id
"431b8083a574440b838df0be431fbe99"
provider_responses
0
endpoint_id
"19b3b5d7-23f6-4b63-8109-64a6cc37d996"
id
"431b8083a574440b838df0be431fbe99"
is_byok
false
latency
1428
model_permaslug
"mistralai/mistral-large-4-0-20261006"
provider_name
"Mistral"
status
200
total_cost
0.03773035
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
0.00%
Matches word count
n/a
neededClean
false
words
549
0.30%
Dialogue to Total Word Ratio
Ratio: 40.79%, Deviation: 10.79%
neededClean
false
wordsTotal
554
wordsDialogue
226
0.1475%