NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M36WQZ3TFNB1H9EA8PCNEBWR
Run Details
Model:
Xiaomi MIMO v2.6 Pro
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
2m 56s
Cost:
$0.010674
Provider:
Xiaomi
Evaluation summary
Score
Evaluator
Details Link
27.17%
Matches word count
(details)
91.20%
Dialogue to Total Word Ratio
(details)
59.1815%
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 unfolded the blueprint across the bonnet of the van and smoothed the creases flat while Bill chewed a pencil stub and squinted at the page like a man trying to read a foreign menu without his glasses. This was the third time she had laid out the garden party map and the third time he had squinted at it without processing a single useful detail. She had learned that Bill required an aggressive preamble before he could operate at full capacity. "Bill. Focus." "I am focused. I'm focused on the fact that this is a blueprint for a bouncy castle." "It's the layout of the garden. The bouncy castle is a key landmark." She jabbed her finger at the red circle near the back of the map. "Gift table. Under the oak tree. We come over the fence, cross the lawn, duck behind the bouncy castle during the balloon animals, and sweep the table." Bill studied the map with the slow, heavy reluctance of a man reading his own prison sentence. The pencil stub had surrendered somewhere around the hydrangea border and lay in two neat pieces on the grass verge. He leaned closer and tapped the swirls near the left margin. "And what's this here? The thing with the swirls?" "The mulch pit. There's a mulch pit by the hydrangeas. The step-over height is about forty centimetres. Trivial." "What could go wrong?" She had prepared a rebuttal involving timing and crowd psychology, but Bill beat her to the punch by simply reading her face and raising both eyebrows with the quiet accusation of a man who had been burned before. "Thirty children, Molly. Thirty. Screaming. Sticky-fingered. Unpredictable. Children are the most dangerous species on the planet." "That's why we strike at the cake-cutting. Every child, adult, and that bloody magician will face the buffet table. Their backs to us." She allowed herself a small, satisfied smile. The plan was elegant. The plan was symmetrical. The plan was almost poetic. Bill rubbed his jaw and stared at the garden beyond the van, where they were parked on a quiet residential street in broad daylight, discussing a felony at volume that would make a public address system blush. A woman with a pram had already walked past twice. "You know what, I've never stolen from children before. I feel like there's a line." "It's not stealing. It's redistributing." "To us." She let the silence do the heavy lifting. Twenty-six wrapped presents. Estimated value close to two thousand pounds. Premium gifting. The mother had booked through a proper planning company, which meant no socks and no craft kits. She had done the arithmetic twice and the haul was genuinely impressive. Bill had stolen from banks, galleries, private collections, and once from a man holding a cricket bat. But he had never stolen from children, and the moral calculus gave him genuine pause. He sat on the tailgate and breathed through his nose. Then he cracked his knuckles with the resigned professionalism of a man who had just accepted his fate. "Right. Four minutes then." "Four minutes. Not a second more."
created_at
"2026-09-23T10:24:52.864Z"
9/23/2026, 10:24:52 AM
model
"xiaomi/mimo-v2.6-pro-20260921"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2629
moderation_latency
(null)
generation_time
175557
tokens_prompt
75
tokens_completion
11011
native_tokens_prompt
71
native_tokens_completion
12234
native_tokens_completion_images
(null)
native_tokens_reasoning
11584
native_tokens_cached
0
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
0.010674465
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.20.0; linux; x64))"
http_referer
(null)
request_id
"req-1790159092-QWgzX9vdWqgybM9AMiMS"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790159092-XuVztiYsacxpqJUnfClB"
upstream_id
"797b2f95-29b9-47cf-a80a-99bc7d9a275c_a21da76a3fba47dd8efb04111ec8b677"
provider_responses
0
endpoint_id
"4354fc36-b8ab-4ced-aa8c-0efb6b4523f1"
id
"797b2f95-29b9-47cf-a80a-99bc7d9a275c_a21da76a3fba47dd8efb04111ec8b677"
is_byok
false
latency
2542
model_permaslug
"xiaomi/mimo-v2.6-pro-20260921"
provider_name
"Xiaomi"
status
200
total_cost
0.010674465
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Xiaomi"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
27.17%
Matches word count
n/a
neededClean
false
words
519
91.20%
Dialogue to Total Word Ratio
Ratio: 31.36%, Deviation: 1.36%
neededClean
false
wordsTotal
523
wordsDialogue
164
59.1815%