NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-06-01-01KT1HN1TW5W4JCJAX67ZBZ562
Run Details
Model:
MiniMax M3
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
10m 19s
Cost:
$0.037233
Provider:
Minimax
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
99.56%
Dialogue to Total Word Ratio
(details)
99.7799%
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
Bill looked up from his sandwich as Molly slid into the booth across from him, her eyes bright with the particular gleam that always preceded disaster. She placed a crayon-drawn map on the table between his plate and the salt shaker, smoothing out a wrinkle with the efficiency of a seasoned criminal. The diner around them hummed with the clatter of forks and the hiss of coffee machines, but Bill could only hear the distant sound of his better judgment packing its bags and slipping out the back door. It showed the entire backyard in loving crayon detail. "I have found us the perfect job," she said. Bill studied the map. It appeared to be drawn by a talented kindergartner on the back of a placemat, complete with stick figures and a surprisingly accurate depiction of a bouncy castle. "If it involves clowns, I'm out." She ignored this completely and launched into her pitch. "Children's birthday party. The Henderson kid. Backyard. Saturday. Presents everywhere. It's a target-rich environment, Bill." He set down his sandwich very slowly, as if any sudden movement might cause the universe to collapse. "Molly. We've talked about this. We said no more kids. We said no more juice boxes." But Molly was already deep in logistics, her finger tracing a path across the crayon fence line. "I went undercover. I have details. Weak fence near the compost heap. The mother will be arguing with a caterer about gluten-free options. The father will be fighting with a bounce house. They'll be helpless. I have the patrol patterns of the dog. I have the WiFi password." Bill pinched the bridge of his nose, a gesture that was becoming his primary form of communication. "And the kids?" "Sugar. Chaos. At least one balloon-related meltdown. It's the perfect distraction." He stared at the map. There was a tiny drawing of a unicorn in the corner, surrounded by what appeared to be musical notes or possibly a very small tornado. He traced the crayon path with a weary finger. "What's my role?" She produced a glossy photograph of a giant mouse costume, complete with a disturbingly cheerful face. "You are the entertainment." Bill looked at the photograph. Then at Molly. Then back at the photograph. Then at the photograph one more time, hoping it would change. "I am not wearing a tail." "You're wearing a tail, Bill. And ears. And whiskers. I've already rented it. Non-refundable deposit." He opened his mouth to protest, but she was already explaining the extraction plan, complete with a modified diaper bag on wheels, a false bottom lined with velvet, and a series of distraction routines involving balloon animals. Bill listened with the defeated expression of a man who had long since accepted his fate. She had clearly rehearsed this particular speech. The olive on his plate rolled slightly as he nodded, slowly, in resigned agreement. "Fine," he said at last, reaching for his coffee. "But I'm keeping the unicorn wrapping paper. And the bow tie."
created_at
"2026-06-01T12:13:34.435Z"
6/1/2026, 12:13:34 PM
model
"minimax/minimax-m3-20260531"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2824
moderation_latency
(null)
generation_time
619013
tokens_prompt
77
tokens_completion
26333
native_tokens_prompt
236
native_tokens_completion
31013
native_tokens_completion_images
(null)
native_tokens_reasoning
25571
native_tokens_cached
222
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
0.03723312
router
(null)
provider_responses
0
endpoint_id
"6851b352-a7e2-45ad-a69e-67ed1208a416"
id
"066caa6e8d7a34847d586beffb8b1551"
is_byok
false
latency
2824
model_permaslug
"minimax/minimax-m3-20260531"
provider_name
"Minimax"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.16.0; linux; x64))"
http_referer
(null)
request_id
"req-1780316014-EZcfRU5Z2lJIHfYY6MAa"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1780316014-5zy9U1HovepBsN7LPSmB"
upstream_id
"066caa6e8d7a34847d586beffb8b1551"
total_cost
0.03723312
cache_discount
0.00010656
upstream_inference_cost
0
provider_name
"Minimax"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
100.00%
Matches word count
n/a
neededClean
false
words
500
99.56%
Dialogue to Total Word Ratio
Ratio: 29.70%, Deviation: 0.30%
neededClean
false
wordsTotal
505
wordsDialogue
150
99.7799%