NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYD68XP2S0WB2NHQ0JH19FGS
Run Details
Model:
MoonshotAI: Kimi K3 (Reasoning, Low)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
31.2s
Cost:
$0.011339
Provider:
Moonshot AI
Evaluation summary
Score
Evaluator
Details Link
2.01%
Matches word count
(details)
1.21%
Dialogue to Total Word Ratio
(details)
1.6115%
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
"We're not monsters," Bill said, tapping the blueprint spread across the hood of the car. "We're opportunists." "It's a six-year-old's birthday party, Bill." "Exactly. Unsecured venue. No cameras. A bouncy castle for cover. It's the easiest score we've ever cased." Molly squinted at the diagram. "You've labeled the gift table 'target zone' and the cake 'secondary objective.'" "The cake is leverage, Molly. Nobody calls the cops when the cake is hostage." She rubbed her temples. "Explain the plan again. Slowly. Like I'm six, which, coincidentally, is the average age of our opposition." Bill straightened up, thrilled. "Phase one: infiltration. You go in dressed as a clown." "Absolutely not." "You own the costume." "That was Halloween, and I was drunk, and I will die before I wear those shoes again." "Fine. You're the magician's assistant." "There is no magician." "There will be. I'm the magician." Bill produced a top hat from nowhere, which, Molly had to admit, was a decent start. "You can't do magic." "I can do the thing where I pull a quarter from behind someone's ear." "That's not magic, that's just touching children. You're going to get arrested before I even get inside." Bill waved her off. "Phase two: the diversion. I gather the kids for a magic show near the bouncy castle. All eyes on me. Meanwhile, you, dressed as a parent—" "Which parent?" "A generic one. Khakis. A polo. Sunglasses indoors. Nobody questions a dad." "I'm a woman, Bill." "A cool aunt, then. A cool aunt loads the presents into the bouncy castle through the emergency flap." "Bouncy castles don't have emergency flaps." "This one will. I made some calls." Molly stared at him. "You commissioned a bouncy castle with a trap door." "I know a guy." "You know a bouncy castle guy?" "He's versatile. He also does ball pits, but those are for bigger jobs." She folded her arms. "Phase three. Tell me there's a phase three that doesn't end in a children's hospital." "Phase three: extraction. I announce the grand finale. The Great Disappearing Presents Trick. Kids lose their minds. Parents applaud. By the time little Timmy realizes his Nerf arsenal is gone, we're doing sixty on the interstate." "And the cake?" "I escape in the confusion. Standard magician exit. Smoke bomb." "You don't have a smoke bomb." "I have a bag of flour and a dream, Molly." Molly pinched the bridge of her nose and breathed deeply. Somewhere in her soul, a small voice asked how her life had come to this. A louder voice asked how much a six-year-old's haul was even worth. "Fine," she said. "But I'm not the cool aunt. I'm the cool uncle. I've got a fake mustache and everything." "Since when?" "Since the museum job. You were unconscious for the good part." Bill beamed and rolled up the blueprint. "Saturday, two o'clock. Bring the mustache. I'll bring the hat." "And if a kid cries?" "We run. That's always been the plan, Molly. We just never had a plan before the running part." "It's a terrible plan." "It's our best plan." "That's the tragedy of it," Molly said, and went to find her khakis.
created_at
"2026-07-25T17:48:23.115Z"
7/25/2026, 5:48:23 PM
model
"moonshotai/kimi-k3-20260715"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2280
moderation_latency
(null)
generation_time
31151
tokens_prompt
75
tokens_completion
815
native_tokens_prompt
147
native_tokens_completion
753
native_tokens_completion_images
(null)
native_tokens_reasoning
17
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
0.0113391
router
(null)
provider_responses
0
endpoint_id
"650c9b12-7b6d-444d-8d92-df8daef84d7a"
id
"chatcmpl-6a64f6e87e9f80111fc5dc20"
is_byok
false
latency
2280
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-1785001703-X8bgDk1Yc6hFvDRxmVcQ"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1785001703-R1MKrXZTbJL1zUbKjezJ"
upstream_id
"chatcmpl-6a64f6e87e9f80111fc5dc20"
total_cost
0.0113391
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
2.01%
Matches word count
n/a
neededClean
false
words
525
1.21%
Dialogue to Total Word Ratio
Ratio: 79.40%, Deviation: 9.40%
neededClean
false
wordsTotal
529
wordsDialogue
420
1.6115%