NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-04-24-01KQ02YTHBQ886DSB9C918GN4T
Run Details
Model:
GPT-OSS 120B
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
48.7s
Cost:
$0.002032
Provider:
WandB
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
50.0007%
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 van idled beneath the maple tree, its engine ticking like a nervous heartbeat. Inside, Molly adjusted the flashlight strap on her chest, while Bill fidgeted with a shiny keychain shaped like a clown. The backyard beyond buzzed with balloons, a bouncy castle, and the distant squeal of giggling children. Molly: 'Okay, Bill, the party starts at six, but the parents will be distracted by the magician at seven. That gives us a ten‑minute window.' Bill: 'Ten minutes sounds tight, but we’ve got the lock‑pick set and the distraction balloon rig I built.' Molly: 'Make sure the rig releases confetti right when we’re inside, so the kids think it’s part of the show.' Bill: 'Got it. I’ll set the timer for 6:58, synchronized with the clown’s balloon animal finale.' Molly: 'And the presents?' Bill: 'All stacked on the table under the rainbow banner. I’ll slip the tarp over them, then we pull the whole thing out like a magician’s curtain.' Molly: 'We need a getaway car. I heard the parents ordered a minivan for the ride home; we can hijack that.' Bill: 'Better than the old sedan that squeaks every time we brake. The minivan has a back door we can open from the inside.' Molly: 'What about the kids? They’ll scream if we’re seen.' Bill: 'We wear clown masks. Nobody will suspect a clown stealing toys.' Molly: 'And the flashlight?' Bill: 'Red LED, low beam, won’t set off any motion sensors. Besides, the party lights are so bright the kids will think it’s a special effect.' Molly: 'Alright, let’s run through the steps one more time.' Bill: 'Step one, park the van two houses down, hide the engine. Step two, slip in through the side gate at 6:58. Step three, deploy confetti rig, grab the tarp, pull the presents. Step four, sprint to the minivan, jump in, and disappear.' Molly: 'Anything I’m missing?' Bill: 'Just remember to leave a balloon animal for the kids; they’ll think it’s a gift and stop looking for the missing ones.' Molly: 'Perfect. Let’s do this.' Molly slipped the flashlight back into its holster, the red glow flickering against her cheek as she checked the side‑door latch. Bill tugged the oversized clown mask from his bag, its oversized red nose wobbling comically. The moon hung low, casting a silvery sheen over the garden gnomes that stared like indifferent judges. A sudden pop of a firecracker from a distant fireworks stand startled a nearby squirrel, which darted across the lawn, its tiny tail a blur. The duo crept forward, their sneakers barely making a sound on the soft grass, the scent of stale cake and frosting hanging in the air. Overhead, a string of fairy lights blinked in rhythm with the distant music, creating a carnival‑like atmosphere that seemed absurd for a crime scene. Molly glanced at Bill, raised an eyebrow, and whispered, “If we’re caught, I’m blaming the clown.” Bill smirked, adjusting his oversized shoes, and replied, “Deal. Now, on three—one, two, three—.”
created_at
"2026-04-24T15:50:26.102Z"
4/24/2026, 3:50:26 PM
model
"openai/gpt-oss-120b"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
168
moderation_latency
(null)
generation_time
48390
tokens_prompt
77
tokens_completion
3338
native_tokens_prompt
125
native_tokens_completion
3356
native_tokens_completion_images
(null)
native_tokens_reasoning
2562
native_tokens_cached
32
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
0.00203235
router
(null)
provider_responses
0
endpoint_id
"3ca6292f-cef7-48a2-bf52-988e934cab57"
id
"chatcmpl-827ff5687c9142d5"
is_byok
false
latency
68
model_permaslug
"openai/gpt-oss-120b"
provider_name
"WandB"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.1; linux; x64))"
http_referer
(null)
request_id
"req-1777045826-Lf55Vug0NU7cxvNQjKHB"
session_id
(null)
api_type
"completions"
id
"gen-1777045826-p4mGtcQluq1d1AaGChnP"
upstream_id
"chatcmpl-827ff5687c9142d5"
total_cost
0.00203235
cache_discount
(null)
upstream_inference_cost
0
provider_name
"WandB"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
100.00%
Matches word count
n/a
neededClean
false
words
500
0.00%
Dialogue to Total Word Ratio
Ratio: 35.00%, Deviation: 15.00%
neededClean
false
wordsTotal
500
wordsDialogue
175
50.0007%