NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-04-24-01KQ03KK6X9GHDP060CZK8H1HJ
Run Details
Model:
Xiaomi MIMO v2.5
Test:
Dialogue tags
Scenario:
Write 200 words with 10% dialogue
Duration:
5.9s
Cost:
$0.000838
Provider:
Xiaomi
Evaluation summary
Score
Evaluator
Details Link
9.61%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
4.8041%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 200 word scene that has 10% 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 unrolled the blueprints across the table, though calling them blueprints was generous—they were crayon drawings she'd lifted from a refrigerator door in apartment 4B. Bill leaned over, squinting at what appeared to be a stick figure guarding a mountain of brightly wrapped boxes. "The target is 42 Maple Drive," Molly whispered, tapping the drawing with a red pen. "Three o'clock sharp. That's when the clown does his naptime routine." Bill raised an eyebrow. "You've done your reconnaissance." "Two weeks of intense surveillance," Molly confirmed. "I attended three birthday parties in that neighborhood. Ate so much cake I nearly blew our cover." "Did you identify security?" "Two dogs. One grandfather. One bouncy castle operator who looked suspicious but turned out to just have a bladder condition." Bill nodded approvingly and studied the drawing further. "What about the noise discipline situation? These events can get loud." "The kids scream on a predictable pattern. Screams at present time, silence during cake, then more screaming when little Timmy opens that plastic drum set." Molly shuddered. "We need to be in and out before the percussion starts." "Getaway vehicle?" "I borrowed a minivan. Puts us right under the radar." Bill straightened up, cracking his knuckles. "This might be our finest hour." "It's definitely our weirdest," Molly admitted, rolling up the blueprints. "Let's go steal Christmas."
created_at
"2026-04-24T16:01:46.729Z"
4/24/2026, 4:01:46 PM
model
"xiaomi/mimo-v2.5-20260422"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1076
moderation_latency
(null)
generation_time
5821
tokens_prompt
77
tokens_completion
480
native_tokens_prompt
311
native_tokens_completion
398
native_tokens_completion_images
(null)
native_tokens_reasoning
92
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
0.00083848
router
(null)
provider_responses
0
endpoint_id
"2af4bf94-26fa-42b9-bce0-8b11071e105a"
id
"2a462a737bbd4958a73a905620b38a83"
is_byok
false
latency
1021
model_permaslug
"xiaomi/mimo-v2.5-20260422"
provider_name
"Xiaomi"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.1; linux; x64))"
http_referer
(null)
request_id
"req-1777046506-4MpjfE5q7rCyh91VhFwn"
session_id
(null)
api_type
"completions"
id
"gen-1777046506-NA5GHXtthjQiSaRdihM5"
upstream_id
"2a462a737bbd4958a73a905620b38a83"
total_cost
0.00083848
cache_discount
0.00008192
upstream_inference_cost
0
provider_name
"Xiaomi"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
9.61%
Matches word count
n/a
neededClean
false
words
222
0.00%
Dialogue to Total Word Ratio
Ratio: 63.06%, Deviation: 53.06%
neededClean
false
wordsTotal
222
wordsDialogue
140
4.8041%