NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYD78QZB81G7DBWQNE8QASN9
Run Details
Model:
MoonshotAI: Kimi K3 (Reasoning, Low)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
27.4s
Cost:
$0.011069
Provider:
Moonshot AI
Evaluation summary
Score
Evaluator
Details Link
2.01%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
1.0058%
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 hand-drawn blueprint across the roof of the getaway hatchback, weighting the corners with two juice boxes and a rubber duck. "The Henderson birthday party. Saturday. One o'clock sharp." Bill squinted at the crayon arrows. "Why is the bouncy castle labeled 'extraction point'?" "Because the kids will be in it, Bill. That's cover. Chaos. Thirty screaming six-year-olds is better than any smoke bomb." "Seems cruel," Bill muttered, adjusting his fake clown nose. "Stealing presents from a child." "We're not stealing from a child. We're stealing from the parents. Half that pile is from a registry. The Hendersons have a trampoline AND a waterslide, Bill. They can take the hit." Molly had done recon for three weeks. She knew the gift table would be staged near the patio doors. She knew the grandmother napped from two to four. She knew the dog, Biscuit, could be neutralized with exactly one and a half hot dogs. "The clown suits bother me," Bill said. "The clown suits are genius. No one looks at the clown. The clown is wallpaper. You could carry a grandfather clock out of there in a clown suit and someone would just say, oh, that's probably part of the act." "I don't know how to make balloon animals." "Then don't. If a kid asks, tell him you're a minimalist clown." Bill practiced his minimalist clown face in the side mirror. It looked mostly like indigestion. Saturday arrived humid and bright. They parked two streets over, waddled through the Hendersons' side gate, and immediately hit their first complication: there was already a clown. "There's already a clown," Bill hissed. "I see that there's already a clown." "Nobody said anything about a clown being booked. He's doing a whole bit. There's juggling." Molly recalculated fast. "Fine. New plan. We're his assistants. Carry his props. Look helpful. Get near the gift table during cake, when every adult in that yard is filming their kid blowing out candles." So they carried props. The real clown, whose name was Giggles and who took his craft with unsettling seriousness, handed Bill a unicycle and told him he had "honest shoes." Molly stationed herself three feet from the presents, cataloguing: a gaming console, a drone, something in a silver box that rattled expensively. Then the candles were lit. Every phone in the county rose in unison. Molly gave the signal, a subtle tug on her left ear. Bill tripped over the unicycle, crashed into the gift table, and came up holding the drone and half a sheet cake. "Run," Molly said. They ran. Through the sprinkler. Past the napping grandmother, who opened one eye, declared "wonderful party," and went back to sleep. Biscuit pursued them for exactly one and a half hot dogs' worth of yard. They dove into the hatchback and peeled out, clown shoes flapping. "How much did we get?" Bill panted. Molly counted. "Four presents." "There were forty presents." "Four is a start. And I got the silver one." She shook it. It rattled expensively. It was a piƱata repair kit. "Next time," Bill said, wiping cake off the steering wheel, "we case the clown situation first."
created_at
"2026-07-25T18:05:45.85Z"
7/25/2026, 6:05:45 PM
model
"moonshotai/kimi-k3-20260715"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2195
moderation_latency
(null)
generation_time
27276
tokens_prompt
75
tokens_completion
809
native_tokens_prompt
147
native_tokens_completion
735
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
0.0110691
router
(null)
provider_responses
0
endpoint_id
"650c9b12-7b6d-444d-8d92-df8daef84d7a"
id
"chatcmpl-6a64fae6e48fcc6c53c42571"
is_byok
false
latency
2195
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-1785002745-dyti9nbUYojkVfFwYFPg"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1785002745-ouNguQZgVc3B0wBVjF4Q"
upstream_id
"chatcmpl-6a64fae6e48fcc6c53c42571"
total_cost
0.0110691
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
0.00%
Dialogue to Total Word Ratio
Ratio: 47.54%, Deviation: 17.54%
neededClean
false
wordsTotal
528
wordsDialogue
251
1.0058%