NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1M00NPRP69DA8HARP4GY95C
Run Details
Model:
Qwen 3.8 27B (Reasoning, XHigh)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
1m 33s
Cost:
$0.026524
Provider:
Phala
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
100.00%
Dialogue to Total Word Ratio
(details)
100.0000%
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 parked the delivery van beside the inflatable castle, where a banner announced the party, and Bill adjusted his paper hat with a serious look of total authority while cupcakes trembled gently inside the house. "Bill, the party has ten tables, five hundred cupcakes, and zero security, which is insulting." Bill pointed at the table, where gifts sat wrapped in bright paper, each one promising joy that could be sold for cash before breakfast, maybe even lunch, and a small tax from the party now. "Zero security means the parents trust us, Molly, which we should absolutely exploit professionally today." Molly pulled a tiny whistle from her apron, because authority, she believed, should always come with sound, especially at parties, where silence was considered rude and therefore useful for serious, professional, absolutely legal schemes tonight. "We enter as bouncers, scan presents, confiscate ribbons, and declare every gift contraband by age." Bill stepped onto the carpet, leaving tiny footprints that looked like question marks, while the party song played too cheerfully in the background and nobody noticed, which was their preferred outcome for everyone present tonight. "No, we enter as judges. The children will line up, and we award prizes publicly." Molly checked her watch, then checked a gift, because time was money and boxes were probably even better, especially if they contained batteries or maybe candy, depending on the year and perhaps small toys too. "Public awards are suspicious. We will be quiet bouncers, using tiny whistles and terrible expressions." Bill smiled, showing one chipped tooth, because he had once judged a pancake contest and still remembered the authority that had made pancakes bow in fear while the syrup watched nervously from the pan below. "Terrible expressions are our specialty. I have practiced mine in every elevator for a decade." Molly tapped a gift with her ring, as if it were a suspicious soup, and nodded once, with professional disappointment because expired presents always tasted like disappointment and she knew that with absolute certainty today. "When a child asks why we keep the presents, say the presents have expired officially." Bill produced stickers from his jacket, tiny stars and rockets that promised redemption, and arranged them on the table with great ceremony because children could be managed with shiny things and careful silence when needed. "If they cry, offer each one a sticker, a promise, and a vague future apology." Molly moved toward the van, her boots squeaking, as the last gift slid into a bag, and the children watched with polite suspicion which was their favorite thing, second only to cake and confetti, always. "After the last gift is in the van, we will be outside before confetti landed." Molly paused by the van, the engine ticking, and considered the evidence of their success, wrapped, bagged, and ready to leave before the parents arrived with cameras and terrible questions that always followed parties ever. "That is not a plan, Molly, that is a birthday crime with excellent pacing tonight."
created_at
"2026-09-03T16:00:47.843Z"
9/3/2026, 4:00:47 PM
model
"qwen/qwen3.8-27b-20260814"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
436
moderation_latency
(null)
generation_time
92797
tokens_prompt
75
tokens_completion
7464
native_tokens_prompt
186
native_tokens_completion
8824
native_tokens_completion_images
(null)
native_tokens_reasoning
8193
native_tokens_cached
64
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
0.026524
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788451247-54GAKquHggGOlKGby4hh"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788451247-9Oe570JHZWBgircW0pvr"
upstream_id
"req_eb7dbd115e297597a734a82a95a76334"
provider_responses
0
endpoint_id
"cfc4d0ca-df7c-4b5f-9c74-328c41a7cbc3"
id
"req_eb7dbd115e297597a734a82a95a76334"
is_byok
false
latency
436
model_permaslug
"qwen/qwen3.8-27b-20260814"
provider_name
"Phala"
status
200
total_cost
0.026524
cache_discount
0.0000224
upstream_inference_cost
0
provider_name
"Phala"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
100.00%
Matches word count
n/a
neededClean
false
words
500
100.00%
Dialogue to Total Word Ratio
Ratio: 30.00%, Deviation: 0.00%
neededClean
false
wordsTotal
500
wordsDialogue
150
100.0000%