NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYD5GSQ0FNYNBXQSATZB6MG7
Run Details
Model:
MoonshotAI: Kimi K3 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
29.9s
Cost:
$0.010589
Provider:
Moonshot AI
Evaluation summary
Score
Evaluator
Details Link
1.04%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
0.5180%
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 Bill's car, weighing down the corners with a half-eaten sandwich and a water pistol. "This," she announced, "is the Hendersons' back garden. Saturday, two o'clock. Twenty-eight confirmed presents." Bill squinted at the drawing. "Why is there a bouncy castle on the blueprint?" "Because there's a bouncy castle at the party, Bill. I do thorough reconnaissance." "It's just that you drew little windows on it. Bouncy castles don't have windows." "Artistic license." Bill picked up the sandwich, took a bite, and put it back on the southeast corner. "So we walk in, grab the presents, walk out. What's the complication?" Molly pulled out a second sheet of paper. "The complication is Timothy Henderson, age seven, birthday boy. According to my source, the kid has wanted a detective kit since Christmas. He's going to be watching that present table like a hawk." "Your source is his babysitter." "She talks a lot when she thinks you're a clown booking agent." "Are you a clown booking agent?" "For the purposes of this conversation, yes." Bill leaned against the car and crossed his arms. "Right. So what's the plan, genius?" Molly tapped the blueprint with a chewed pencil. "Phase one: we arrive disguised as the entertainment. You're the magician, I'm your lovely assistant. Phase two: you perform tricks so bad that every adult in the garden leaves to go 'check on the barbecue.' Phase three: I sweep the present table into a decorative sack." "Why am I the magician? I don't know any magic." "That's what makes it convincing, Bill. Nobody suspects a magician who is genuinely terrible at magic." "What about the kid detective?" Molly smiled the smile she reserved for safes, salesmen, and small obstacles. "Phase four. I hire him." "You hire him." "Junior detectives can't resist a case. I tell him someone stole my lucky pencil, he spends three hours interrogating the clown from next door's party, we walk out with a mountain of Lego." Bill was quiet for a moment, staring at the blueprint, the sandwich, and the water pistol. "Molly, can I ask you something?" "Go on." "We've robbed banks. We've robbed a casino. We once stole an entire boat. Why are we robbing a children's birthday party?" Molly rolled up the blueprint and tucked it under her arm. "Because, Bill, the Hendersons' insurance payout on those presents is laundered through their garden gnome business, and the father shorted me forty quid on a ceramic fisherman in 2019." "So it's revenge." "It's revenge with presents. The best kind." Bill picked up the water pistol and weighed it thoughtfully. "Do I get a wand?" "A wand and a cape." "A cape that swirls?" "A cape that swirls magnificently." He tucked the pistol into his waistband like a tiny, soggy sidearm. "Then we've got a party to crash."
created_at
"2026-07-25T17:35:12.615Z"
7/25/2026, 5:35:12 PM
model
"moonshotai/kimi-k3-20260715"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2513
moderation_latency
(null)
generation_time
29853
tokens_prompt
75
tokens_completion
770
native_tokens_prompt
147
native_tokens_completion
703
native_tokens_completion_images
(null)
native_tokens_reasoning
47
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
0.0105891
router
(null)
provider_responses
0
endpoint_id
"650c9b12-7b6d-444d-8d92-df8daef84d7a"
id
"chatcmpl-6a64f3d10f3382baf2a9f3e2"
is_byok
false
latency
2513
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-1785000912-hTN0dlbV0deEa8Ok50wn"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1785000912-Cot6nPlr8R3wMzbb5S4j"
upstream_id
"chatcmpl-6a64f3d10f3382baf2a9f3e2"
total_cost
0.0105891
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
1.04%
Matches word count
n/a
neededClean
false
words
474
0.00%
Dialogue to Total Word Ratio
Ratio: 72.27%, Deviation: 42.27%
neededClean
false
wordsTotal
476
wordsDialogue
344
0.5180%