NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-04-24-01KQ022YZWDAGGTE8ECK9A3QT6
Run Details
Model:
Xiaomi MIMO v2.5 Pro
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
25.0s
Cost:
$0.003203
Provider:
Xiaomi
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.03%
Dialogue to Total Word Ratio
(details)
0.0134%
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
Molly tapped the crudely drawn map of the community center with a sparkly purple pen she’d “borrowed” from the party-planning office. “Alright, Operation Pinata Plunder is a go. Step one: you, Bill, distract the clown.” Bill squinted at the floor plan. “The clown? Why not the magician? Magicians are scarier. They pull live birds from their sleeves. That’s unnatural.” “The clown is our central security risk. He’s got a bad knee, loves to talk about his ‘glory days’ juggling for medieval reenactment fairs. You’re a natural conversation partner.” “I was a medieval studies minor.” “Exactly. You’ll keep him occupied with a passionate debate on trebuchet engineering. Meanwhile, I’ll be over here,” she circled a square labeled “PRESENT TABLE (UNATTENDED - PARENTS FOCUSED ON MELTING ICE SCULPTURE),” “liberating the loot. We’ll be in and out in twelve minutes.” “Twelve minutes? That’s ambitious. What if there’s a ‘Pin the Tail on the Donkey’ crisis? Or a sugar rush?” “We adapt. Improvise. But we stick to the timeline.” Molly smoothed out the napkin blueprint. “What’s our exit strategy?” “We don’t talk about exit strategy.” Bill sighed, running a hand through his hair. “I still think we’re targeting the wrong demographic. The presents at a ten-year-old’s party are all… toys. Nerf guns. Video games. What’s the take? Two hundred bucks, tops?” “Are you planning on keeping them, you sentimental fool?” Molly asked, her voice dripping with theatrical scorn. “We’re professionals. We’ll fence the lot to that guy, Darren, who runs the ‘Retro Toy Emporium’ in the strip mall. He’s always looking for pristine, unopened stock. His clientele is a bunch of nostalgic thirty-somethings with disposable income. We’re providing a service, really.” “You’re a philanthropist of crime, Molly. Truly.” “Mock all you want. But when you’re sipping a piña colada on a deck somewhere warm, paid for by a resold Lego Millennium Falcon, you’ll thank me.” Bill peered out the window of their stolen van, watching the streamers flutter on the community center’s front door. “I still feel a bit… low. Stealing from kids. It’s not exactly Ocean’s Eleven, is it? It’s more… Ocean’s eleven-minus-years-of-life-experience.” “It’s not about the kids, Bill. It’s about the challenge. The stealth. The sheer audacity of walking out with a sack of presents while everyone’s watching a hired princess sing ‘Let It Go’ for the fourth time. This is where legends are born.” “Legends,” Bill repeated flatly. “So, after the clown, what’s my next move?” “You hover near the cake table. Look suspicious. When the mom in charge of slices gives you the stink-eye, use the code phrase.” “Right. ‘Is that buttercream or fondant?’” “No, that’s the phase-two code. The distraction code is ‘Has anyone seen my niece, Zoe? She’s wearing a purple cape.’ Then you wander off, drawing their concern, and potentially the entire organizing committee, towards the empty craft room.” “And then?” “And then you meet me by the back emergency exit. You’ll have the gift bags. I’ll have the backpack full of the premium haul from the larger boxes. We blend into the panic when the ice sculpture finally collapses, which based on this floor plan and the location of the buffet, is imminent.” Bill took a deep breath, the crinkled map crinkling in his nervous grasp. “Alright. We’re doing this. We’re heisting a children’s party.” Molly clapped him on the shoulder, a grin spreading across her face. “Trust the plan, Bill. Now, go memorize the signs of a poorly constructed trebuchet. Your clown awaits.”
created_at
"2026-04-24T15:35:13.159Z"
4/24/2026, 3:35:13 PM
model
"xiaomi/mimo-v2.5-pro-20260422"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1651
moderation_latency
(null)
generation_time
24472
tokens_prompt
77
tokens_completion
1226
native_tokens_prompt
315
native_tokens_completion
1031
native_tokens_completion_images
(null)
native_tokens_reasoning
179
native_tokens_cached
256
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"
usage
0.0032032
router
(null)
provider_responses
0
endpoint_id
"5c24ecc9-5ba1-4af0-8877-3402889e7d0a"
id
"f9e1106392294ea28a89d7a972861e5a"
is_byok
false
latency
1567
model_permaslug
"xiaomi/mimo-v2.5-pro-20260422"
provider_name
"Xiaomi"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.1; linux; x64))"
http_referer
(null)
request_id
"req-1777044913-cWwGPUKyXfsEXQ6NA6TU"
session_id
(null)
api_type
"completions"
id
"gen-1777044913-dKVC7YPPPFjYuD1ELpV7"
upstream_id
"f9e1106392294ea28a89d7a972861e5a"
total_cost
0.0032032
cache_discount
0.0002048
upstream_inference_cost
0
provider_name
"Xiaomi"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
577
0.03%
Dialogue to Total Word Ratio
Ratio: 82.82%, Deviation: 12.82%
neededClean
false
wordsTotal
588
wordsDialogue
487
0.0134%