NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-11-01KKFCWR73NWMNEX15D3J9SSSB
Run Details
Model:
ByteDance Seed 2.0 Lite
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
1m 27s
Cost:
$0.006903
Provider:
Seed
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.29%
Dialogue to Total Word Ratio
(details)
0.1442%
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
Crouched behind an overgrown spiraea hedge, Molly and Bill stared across a sun-dappled suburban lawn at a seventh birthday party in full swing. Neon streamers snapped between oak trees, a T-Rex piñata swayed above a screaming huddle of seven-year-olds, and a stack of wrapped presents towered on the back deck. Bill’s dollar-store clown mask sat lopsided on his head, one eye hole gaping over his cheek, and he sweated through a ketchup-stained pizza delivery uniform borrowed from his cousin. Molly clutched a crumpled list and a “World’s Best Cat Mom” canvas tote, a cheap toy grappling hook forgotten at her feet. The party host was inside refilling punch, only a golden retriever puppy standing guard between them and the loot. “If you knock over one more lawn gnome, I’m leaving you here to be the party’s entertainment. Mrs. Lopez over there already waved at you twice, thinks you’re the princess clown she booked for the afternoon,” Molly hissed, jabbing Bill hard in the ribs with her elbow. “This mask was supposed to be my disguise! How was I supposed to know the eye holes are sized for a kindergartener? I can barely make out the present pile, let alone sneak across a lawn full of sprinkler puddles. Also, why are we stealing kids’ birthday presents again? I thought we quit small-time grifts to boost jet skis from lake houses. This feels mean,” Bill mumbled, yanking the mask up far enough to see straight. “We’re not stealing them to resell, you walnut. That My Little Pony castle on top of the pile is for my niece, whose mom refused to buy it because ‘toys promote consumerism.’ The Lego Star Wars set is for my foster kid cousin who’s never had a big birthday, and the Hatchimals are for the little boy down the street whose mom lost her job and couldn’t throw him a party. And if you’re gonna whine about morality, remember last month you tried to steal a grandma’s ceramic cat casserole dish from a church bake sale because you thought it was solid silver,” Molly said, checking items off her crumpled list. “Wait, we’re Robin Hood for first graders? That’s the grift? What’s the actual heist plan, anyway? I can’t climb that deck railing in this getup, I tripped over a fire hydrant on the walk over and scraped my knee,” Bill said. “Simple. I’ll pretend I’m the local librarian who dropped off craft supplies earlier, wander over to leave a birthday card, stuff all the presents into this tote, and you distract that puppy by the deck with a slice of your leftover pepperoni pizza. Easy. If you pull it off, you can have the frosted unicorn cupcake sitting right next to the presents,” Molly said, adjusting a fake library lanyard she’d printed off at Walgreens that morning. “Deal. Wait, that’s it? I wore a clown mask for a plan that’s just lying to a suburban mom?” Bill said, grabbing the pizza slice from his vest pocket. Before he could move, a pigtailed seven-year-old ran past the hedge, spotted his half-pulled-on mask, and screamed “CLOWN!” at the top of her lungs. Half the party turned to stare, and Molly shoved Bill flat into the hedge’s prickly branches, whispering, “You blew it. Now you better learn how to make balloon animals, fast.”
created_at
"2026-03-11T21:29:53.134Z"
3/11/2026, 9:29:53 PM
model
"bytedance-seed/seed-2.0-lite-20260309"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
421
moderation_latency
(null)
generation_time
87120
tokens_prompt
77
tokens_completion
3748
native_tokens_prompt
115
native_tokens_completion
3437
native_tokens_completion_images
(null)
native_tokens_reasoning
2700
native_tokens_cached
0
num_media_prompt
(null)
num_input_audio_prompt
(null)
num_media_completion
0
num_search_results
(null)
origin
"https://ncbench.com/"
is_byok
false
finish_reason
"stop"
native_finish_reason
"stop"
usage
0.00690275
router
(null)
provider_responses
0
endpoint_id
"bf1a3473-7a0a-46d3-aaa5-adb8f876a697"
id
"021773264593250e9db803be2da2d1fb0370cddd09ba5d6dd2248"
is_byok
false
latency
421
model_permaslug
"bytedance-seed/seed-2.0-lite-20260309"
provider_name
"Seed"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.0; linux; x64))"
http_referer
(null)
api_type
"completions"
id
"gen-1773264593-dKAyj7pwpaHR8aFGCy0I"
upstream_id
"021773264593250e9db803be2da2d1fb0370cddd09ba5d6dd2248"
total_cost
0.00690275
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Seed"
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
554
0.29%
Dialogue to Total Word Ratio
Ratio: 60.82%, Deviation: 10.82%
neededClean
false
wordsTotal
564
wordsDialogue
343
0.1442%