NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-06-01-01KT1GREKWTBRD3XV01YZR739C
Run Details
Model:
MiniMax M3
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
5m 25s
Cost:
$0.017447
Provider:
Minimax
Evaluation summary
Score
Evaluator
Details Link
1.04%
Matches word count
(details)
98.09%
Dialogue to Total Word Ratio
(details)
49.5617%
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 blueprints across the hood of their getaway car, which was currently disguised as a clown-mobile for reasons she refused to explain. Bill chewed a toothpick and stared at the schematics like they owed him money. The Sugarplum Birthday Extravaganza. Two o'clock. Twelve kids. One piƱata. Forty-three presents. Molly tapped the paper with a glittery pen, her eyes gleaming with the kind of maniacal focus that had gotten them out of three countries and into fourteen awkward situations. She had circled the gift table three times in red marker, and Bill had learned not to ask why. "Security?" Bill asked. "What security? There's a man in a dog suit. Inflatable. He goes by Captain Wags." She slapped a manila folder onto the hood. "Peanuts, belly rubs, abandonment issues." Molly had a dossier on everything. Bill had once seen her compile a twenty-page report on a suspicious pigeon. He didn't ask questions anymore. The pigeon, in fairness, had been making eye contact for forty-five minutes. "Entry points," she continued, circling the back door. "Kitchen window or we tunnel through the bouncy castle." "Real options?" "Except the bouncy castle. Reinforced." She smiled. "At 2:07 the mom checks the cake. Seven-minute window." Bill nodded. Seven minutes was an eternity in heist time. He could never tell with Molly. Her internal clock ran on caffeine and spite, and she had once pulled off a museum job in four minutes flat because the gift shop had a sale on scarves. "You distract Captain Wags. I grab the loot. Rendezvous at the bouncy castle, exit through the garage." She reached into her bag and produced a squeaky toy shaped like a bacon strip. "This should do it." "You bought him a present." "I bought him an accomplice. By the time he realizes the bacon was a ruse, we'll be halfway to Hector." Hector was their fence, a man who had once asked so few questions he'd made it into Molly's will. Bill had stopped reading her will after the third paragraph, somewhere around "and to my beloved cactus." "Molly, are you okay?" "Thriving, Bill. Professional to my core. Think of the score. Beanie Babies. Hamster balls." "We're going to hell." "We're going to the suburbs. Same difference." She tossed him the squeaky bacon. "Costumes. Caterer? Janitor? Disguised as another dog?" "How about I just wear a hat." Molly blinked. Then she grinned, the kind of grin that preceded either brilliance or a stint in federal custody. There was rarely a middle ground with Molly. Her moral compass had been replaced with a moral protractor, and even that was bent. "Nobody questions a hat. Get in the clown car." Bill climbed into the clown car, which still smelled faintly of seltzer and regret. Molly started the engine, and somewhere in the distance, a child screamed with joy. Bill prayed it wasn't because of them. It was definitely because of them. Somewhere, a mother was about to discover that the gift table was lighter than expected, and a man in an inflatable dog suit was about to have the worst day of his poorly ventilated life. "Now," she said, "who wants cake?"
created_at
"2026-06-01T11:57:57.248Z"
6/1/2026, 11:57:57 AM
model
"minimax/minimax-m3-20260531"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1184
moderation_latency
(null)
generation_time
324087
tokens_prompt
77
tokens_completion
14839
native_tokens_prompt
236
native_tokens_completion
14503
native_tokens_completion_images
(null)
native_tokens_reasoning
14026
native_tokens_cached
114
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
0.01744704
router
(null)
provider_responses
0
endpoint_id
"6851b352-a7e2-45ad-a69e-67ed1208a416"
id
"066ca6c5874cb20aa56d4d9f4a79e79a"
is_byok
false
latency
1184
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-1780315077-EhDeoT7s7sU5VqUGNNlc"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1780315077-ZVHnrWgcVVqMPO2DaNzO"
upstream_id
"066ca6c5874cb20aa56d4d9f4a79e79a"
total_cost
0.01744704
cache_discount
0.00005472
upstream_inference_cost
0
provider_name
"Minimax"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
1.04%
Matches word count
n/a
neededClean
false
words
526
98.09%
Dialogue to Total Word Ratio
Ratio: 29.38%, Deviation: 0.62%
neededClean
false
wordsTotal
531
wordsDialogue
156
49.5617%