NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-10-07-01M4AVASGC3JQ0F35SX0PK502R
Run Details
Model:
Mistral Large 4.0 (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
4m 33s
Cost:
$0.018665
Provider:
Mistral
Evaluation summary
Score
Evaluator
Details Link
20.19%
Matches word count
(details)
98.79%
Dialogue to Total Word Ratio
(details)
59.4923%
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
They sat in Bill's rust-eaten van, parked three blocks from the community center where seven-year-olds were ingesting alarming quantities of sugar. Molly spread a laminated floor plan across her knees, squinting at the bounce castle. Outside, a red balloon escaped its tether and drifted toward the power lines. The GPS chimed. They were almost out of time. Bill checked his watch, then the clown nose. "Okay," Molly said, tapping the blueprint. "The cake table is here, the gift table is here, and the children are everywhere. How do we avoid them?" "Disguise," Bill said, pulling a clown nose from his pocket. "Simple. The birthday boy won't suspect a thing." "That's a felony with face paint. Besides, you're deathly afraid of balloons." "I'm strategically cautious around pressurized latex." "Bill, we're stealing presents from a seven-year-old's party. Do you understand the optics?" "Optics? Molly, it's target rich. LEGO sets. Gaming consoles. Have you seen resale value on Pokémon cards?" "Actual children will be crying." "They'll cry anyway at the sugar crash. We're expediting the inevitable." "You're a monster." "I'm a businessman. We go in as magician's assistants. I juggle, you distract with a puppet, we load the van." "And if the puppet fails?" "Then I juggle harder. I've been practicing with cantaloupes." "There will be parents." "Parents are soft targets. Distracted by wine and sciatica complaints. She'll be on the floor in minutes." "Wait, did you actually research the parents?" "Every last one. Gary from accounting has a bad knee. The mother in blue is lactose intolerant. We strike during the juice box distribution." "That's oddly specific." "I do research. We rob dreams. And also LEGO." "We're committing grand larceny in the presence of minors." "It's the same thing poetically. What's the getaway? I thought the ball pit." "Three feet deep, filled with screaming toddlers. You'd drown." "I'd blend in. I have the emotional maturity of a nine-year-old." "That explains our last job." "Which one? The aquarium? The library? The petting zoo riot?" "All of them. Are you sure about this? The moral weight..." "Is crushing my soul, but think of the profit. That kid is turning seven. Seven! Do you know how many presents accumulate? It's a glitter-covered goldmine." Molly rubbed her temples. The clown nose squeaked as Bill fidgeted with it, stretching the rubber until it resembled a deflated sausage. Through the grimy window, parents arrived with balloons, already looking exhausted. "Fine. But if I wear a tutu, I'm billing you for therapy." "Already budgeted. Line item: 'Psychological Damages - Clown Related.'" "Get that nose out of my sightline." "It's not even on." "Put it away before I use it as a weapon." "We're thirty seconds out. Ready to make history? To become the most hated criminals in the county?" "Ready to be on a watchlist." "Same thing. Let's go steal some childhoods." "Literally." "Semantics." Bill grabbed the wheel. The van lurched forward, tires squealing. Somewhere in the distance, a child began to cry, followed by a juice box hitting the floor. Molly didn't look back. She clutched the floor plan and prayed the therapy bills wouldn't exceed the LEGO profits.
created_at
"2026-10-07T09:32:52.117Z"
10/7/2026, 9:32:52 AM
model
"mistralai/mistral-large-4-0-20261006"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3109
moderation_latency
(null)
generation_time
272871
tokens_prompt
75
tokens_completion
8300
native_tokens_prompt
67
native_tokens_completion
8909
native_tokens_completion_images
(null)
native_tokens_reasoning
7465
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
0.01866537
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1791365572-Yi6OfttXXIM0tGRdkfiL"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1791365572-RY0zpXOW5Wmw0EkWDmqZ"
upstream_id
"a4e2df7c843040718843222a87b9a59e"
provider_responses
0
endpoint_id
"19b3b5d7-23f6-4b63-8109-64a6cc37d996"
id
"a4e2df7c843040718843222a87b9a59e"
is_byok
false
latency
721
model_permaslug
"mistralai/mistral-large-4-0-20261006"
provider_name
"Mistral"
status
200
total_cost
0.01866537
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
20.19%
Matches word count
n/a
neededClean
false
words
520
98.79%
Dialogue to Total Word Ratio
Ratio: 69.51%, Deviation: 0.49%
neededClean
false
wordsTotal
528
wordsDialogue
367
59.4923%