NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1M0Q8SBJ63RVH22KY9M81F3
Run Details
Model:
Qwen 3.8 27B (Reasoning, XHigh)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
4m 59s
Cost:
$0.042754
Provider:
Reka
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 50% 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
The basement smelled of popcorn and wet feet while the party thumped above like a tiny terrible drum and balloons popped all last night again. "Bill, the gift table is behind that inflatable castle, and I need you to be useful while I check the cake and frosting and glitter." Bill nodded and adjusted his paper crown, which was lopsided, crayon stained, and somehow still very official under the spotlight with dignity too on site. "I am useful, Molly, because I can open wrappers without making a single embarrassing noise while you handle the shiny loot and prizes every time." They crouched behind a table covered in streamers that looked like leftovers from a confused, dripping, glittery rainbow stuck to the floor and walls nearby. "My plan is simple enough. We take all the presents, slip out the window, let the piñata blame itself because it already looks guilty today." Molly tapped the window with one gloved finger, pretending the glass was a vault that had bad manners and would not open for us now. "Bad manners? Bill, you once called a birthday clown a security system in a tight suit and I respect his work a lot too really." Above them, a child shrieked with delight, and Bill immediately whispered that the guard had changed routes for the third time this evening again now. "There is no guard. There is only a toddler with a juice box and zero respect for property while it honks its tiny horn today." Molly held up one finger, and Bill froze, even though the only danger was a cup of cake full of suspicious sprinkles and sugar dust. "That cup is booby trapped, I can feel it. The sprinkles are arranged like tiny surveillance cameras, especially the red ones and the green one." Bill leaned forward, his ears pink, and announced that he could hear the birthday songs upstairs singing about a plastic dinosaur named Gary and friends. "Gary is our decoy. We bribed Gary's owner using a stolen cookie and a very serious promise to protect the presents tonight, very carefully indeed." Molly grinned, then pulled a tiny ladder from inside a cardboard box shaped like a rocket ship and a paper star that sparkled brightly outside. "A rocket ship box? Molly, when did you smuggle that in, and why is it so perfectly useful for our present heist tonight right now." Bill stared at the ladder and said nothing, which was his most honest, dangerous look like a secret agent who had forgotten something crucial always. "You are not a secret agent, Bill. You are a gift thief wearing a paper crown, and that is your title for the rest always." They shared a look that promised trouble, candy theft, and one very confused piñata by morning and a missing gift table nearby again for sure. "Then we go now, because the presents are waiting, and I refuse to share with children, especially the loud ones in the party tonight here."
created_at
"2026-09-03T16:13:08.274Z"
9/3/2026, 4:13:08 PM
model
"qwen/qwen3.8-27b-20260814"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
305
moderation_latency
(null)
generation_time
299397
tokens_prompt
75
tokens_completion
13501
native_tokens_prompt
118
native_tokens_completion
16755
native_tokens_completion_images
(null)
native_tokens_reasoning
12790
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
0.04275357
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788451988-YWsee3xIUrXY5A2AVFDk"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788451988-WPIl9APFNexGbr2k9qVK"
upstream_id
"chatcmpl-7b4c4c40-21ec-4cda-8149-934d24ab43a3"
provider_responses
0
endpoint_id
"15086968-df2f-4d69-86bb-c74c31644bb6"
id
"chatcmpl-7b4c4c40-21ec-4cda-8149-934d24ab43a3"
is_byok
false
latency
303
model_permaslug
"qwen/qwen3.8-27b-20260814"
provider_name
"Reka"
status
200
total_cost
0.04275357
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Reka"
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: 50.00%, Deviation: 0.00%
neededClean
false
wordsTotal
500
wordsDialogue
250
100.0000%