NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1KY5PF6KKGHZWQE3JN74SBJ
Run Details
Model:
Qwen 3.8 27B (Reasoning, XHigh)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
3m 55s
Cost:
$0.039436
Provider:
Io Net
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 and Bill crouched outside the bouncing castle, their fingers itching for ribbon, tissue, and small boxed miracles. The children's party was a glittering vault of innocence, and every present inside had their name in crayon. Tonight, the party would become their most delightful heist, wrapped in bubblegum and confusion. "Bill, I want every gift under twelve pounds, but if a dragon truck appears, we take the dragon truck, because our van already smells like birthday candles. The kids will cry, the parents will blame the clown, and we will be gone before the cake is sliced into the night." Bill adjusted his party hat, which had been stolen from a previous job, and checked the inflatable slide for weak seams, soft padding, and easy escape routes. He tapped the wall twice, grinned at Molly, and counted the juice boxes stacked like shiny gold bars on the counter right there. "Molly, I can crawl through the punch table, but only if we leave the cake behind, because one bite ends professionalism, and expensive frosting is definitely guarded by a parent holding a clipboard and terrible authority. We can take presents without tasting dessert, which protects our reputation every single time." Molly smiled, because Bill's fear of frosting was useful. She slipped her hand through the tent flap and felt the warm, sugary air of the party pressing against her skin. It smelled like glitter, popcorn, and presents that had already been wrapped with suspiciously poor taste and zero security tonight. "Bill, count the gifts by ribbon color, blue for trucks, pink for dolls, purple for mysterious boxes, and yellow for anything that might beep or explode. When the piƱata bangs, we run, because a swinging donkey is not a security guard, but it can still make a mess right now." Bill slid under the punch table, his knees squeaking like a bad zipper, and his elbows brushed a mountain of cupcakes that trembled but did not fall. From that vantage point, every present looked like a jewel, and every balloon looked like a witness trying to float away at once. Molly stepped inside, wearing a stolen superhero cape that made her look less like a criminal and more like a disappointed librarian who had chosen costuming. She gathered a rocket, a stuffed unicorn, and a labelless box, which felt dangerously important and possibly alive under the table in the dark. Bill stacked gifts into a laundry basket, while Molly fed them through the tent opening like a polite conveyor belt, and the clown watched from across the room, frowning at a balloon animal. The box hummed, so Molly judged the moment perfect for an urgent, glittery departure before dessert arrived. They burst out, capes fluttering, laundry basket full, and the dragon truck squealing behind them on wheels that had not been approved for theft. By the time the parents noticed, the only evidence left was a party hat, a sticky trail, one small ribbon, and a very confused clown outside.
created_at
"2026-09-03T15:28:35.31Z"
9/3/2026, 3:28:35 PM
model
"qwen/qwen3.8-27b-20260814"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
427
moderation_latency
(null)
generation_time
234904
tokens_prompt
75
tokens_completion
11416
native_tokens_prompt
118
native_tokens_completion
12871
native_tokens_completion_images
(null)
native_tokens_reasoning
10668
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
0.039436236
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788449315-O2DbLvZhCsV4VHzpjvkX"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788449315-rQ8ayS9p9cHERx6aezb7"
upstream_id
"chatcmpl-b1fe95a6-385d-4b89-b0da-300ca6e975c1"
provider_responses
0
endpoint_id
"c65c4547-63a7-49fb-8354-a45d118663c3"
id
"chatcmpl-b1fe95a6-385d-4b89-b0da-300ca6e975c1"
is_byok
false
latency
426
model_permaslug
"qwen/qwen3.8-27b-20260814"
provider_name
"Io Net"
status
200
total_cost
0.039436236
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Io Net"
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%