NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYD61Q1T8FY9WQQ50WFY08RR
Run Details
Model:
MoonshotAI: Kimi K3 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
31.3s
Cost:
$0.012374
Provider:
Moonshot AI
Evaluation summary
Score
Evaluator
Details Link
68.10%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
34.0511%
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 blueprint across the hood of their getaway hatchback, weighting the corners with a juice box and a rubber duck Bill had refused to explain. "Now pay attention," she said. "The Peterson kid turns six on Saturday. House full of presents. In and out, eleven minutes." "Eleven minutes," Bill repeated, writing it on his palm. "Wait, why eleven?" "Because twelve is showing off." Bill nodded like this was ancient wisdom. He was six foot three, built like a wardrobe, and had once gotten stuck in a dog door during what was supposed to be a routine burglary of a house with no dog. Molly kept him around because he was loyal, and because he could carry a grandfather clock without complaining. She traced the route with a crayon, since it was the only writing utensil in the car. "Front lawn. Bounce house here. A magician here. The presents go on a table on the patio, right there." "Who hires a magician?" Bill asked, genuinely alarmed. "Magicians have capes, Molly. You can't see what their hands are doing." "Their hands are doing card tricks, Bill." "That's what they want you to think." Molly took a breath the way you do when you're explaining to your accountant why you listed 'loot' as a business expense. "Here's the plan. We arrive at two o'clock dressed as the entertainment." "What entertainment?" "Clowns." Bill's face went pale beneath the layer of sunscreen he wore on every job, criminal or otherwise. "Molly, no." "Bill, yes." "Molly, I'm terrified of clowns." "You ARE the clown. That's the whole point. Nobody suspects the clown." "I'd suspect me." She handed him the costume. He held it at arm's length like a dead raccoon. The wig alone had clearly witnessed things. "While the kids are distracted by cake," Molly continued, "we move to the patio. We bag the presents. Out the side gate, into the car, gone before the piñata breaks." Bill studied the blueprint. "What's this circle here?" "That's the piñata." "No, the other circle. With the teeth." Molly squinted. "That's the mother-in-law. She watches the gift table like a hawk with opinions." "Can we bribe her?" "With what?" "I've got a nice fruitcake in the car." "You are not giving a fruitcake to a grandmother at a six-year-old's birthday party." "It's got brandy in it." Molly paused. "Keep the fruitcake handy." Bill smiled the smile of a man who had contributed something for once. He rolled up the blueprint and tucked it into his overalls, where it sat next to, as far as Molly could tell, another rubber duck. "One more thing," she said. "If anything goes wrong, the fallback is the bounce house." "Escape through the bounce house?" "Nobody chases anyone through a bounce house, Bill. It's physics." Bill nodded slowly, writing 'physics' on his other palm, right below where he'd written 'eleven minutes' and, for reasons unknown, the word 'soup.'
created_at
"2026-07-25T17:44:26.945Z"
7/25/2026, 5:44:26 PM
model
"moonshotai/kimi-k3-20260715"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2363
moderation_latency
(null)
generation_time
31196
tokens_prompt
75
tokens_completion
899
native_tokens_prompt
147
native_tokens_completion
822
native_tokens_completion_images
(null)
native_tokens_reasoning
141
native_tokens_cached
147
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
(null)
usage
0.0123741
router
(null)
provider_responses
0
endpoint_id
"650c9b12-7b6d-444d-8d92-df8daef84d7a"
id
"chatcmpl-6a64f5fcb1a1291b1d65ab7b"
is_byok
false
latency
2363
model_permaslug
"moonshotai/kimi-k3-20260715"
provider_name
"Moonshot AI"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.18.0; linux; x64))"
http_referer
(null)
request_id
"req-1785001466-CCeI7Ydorx87bgw4L9MS"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1785001466-6aeWMZ5brK8YJ3eSImdu"
upstream_id
"chatcmpl-6a64f5fcb1a1291b1d65ab7b"
total_cost
0.0123741
cache_discount
0.0003969
upstream_inference_cost
0
provider_name
"Moonshot AI"
response_cache_source_id
(null)
data_region
"global"
Evaluation details
Result
Evaluator
Details
Meta Data
68.10%
Matches word count
n/a
neededClean
false
words
486
0.00%
Dialogue to Total Word Ratio
Ratio: 54.08%, Deviation: 24.08%
neededClean
false
wordsTotal
490
wordsDialogue
265
34.0511%