NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-04-24-01KQ0AZN0RDM41S8HK3YT6QTBQ
Run Details
Model:
GPT-OSS 120B
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
28.2s
Cost:
$0.000268
Provider:
Novita
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
20.26%
Dialogue to Total Word Ratio
(details)
10.1304%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 70% 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 night air was thick with the smell of cake frosting drifting from the McAllister residence, where a sea of balloon animals and glittering streamers promised a loot like no other. Molly crouched behind the hedge, her neon sneakers squeaking against the mulch as she adjusted the night‑vision goggles she’d won in a poker game. “Bill, you sure this is the right place?” she whispered, eyes scanning the backyard where a plastic castle loomed under a canopy of rainbow ribbons. “Positive,” Bill replied, sliding his lockpick set from his pocket like a magician pulling a rabbit from a hat. “The kids are out, the parents are distracted by the karaoke version of ‘Baby Shark,’ and the presents are piled higher than my mother’s expectations at a high school graduation.” Molly snorted. “You always compare heists to family events. It’s charming, really.” Bill smirked. “Only because my mother’s expectations are the only thing that ever scares me.” “Speaking of scares,” Molly said, tapping the side of her earpiece, “the security camera feeds show the host’s aunt—she thinks she’s a ninja, but really she’s just a 57‑year‑old with a bob haircut and a pacifier‑shaped keychain. She’s patrolling the garden every ten minutes.” “Perfect,” Bill said, flipping a tiny screwdriver like a baton. “She’ll be the perfect distraction. I’ll plant the smoke bomb in the punch bowl. By the time she notices a slightly caramel‑scented fog, we’ll be making our grand exit with the gift table.” Molly raised an eyebrow. “You want to set off a smoke bomb in a punch bowl? That’s either genius or a recipe for a sweet‑tooth disaster.” “It’s all about the flavor profile,” Bill said, winking. “Peanut butter and jelly. All the kids will think it’s a new drink trend.” Molly laughed, the sound muffled by the thick leaves. “Alright, what’s the plan for the presents themselves? Are we grabbing everything or just the high‑value items?” “First, the Lego sets. Those are worth more than my credibility as a respectable citizen,” Bill replied, tapping his nose. “Then the video game consoles—if we can get past the protective bubble wrap. Finally, the gold‑wrapped chocolate coins. They’re literally edible money.” Molly pretended to think. “Do we have a way to get past the bubble wrap without triggering the alarm?” “Yes,” Bill said, pulling out a roll of tape. “I’ll apply it to the underside of the table. When we lift the box, the tape will keep the wrapper from moving, and the pressure sensor will think nothing’s happening.” Molly glanced at the clock on the wall of the house, its hands pointing to 8:03 p.m. “We’ve got exactly fourteen minutes before the clown—what’s his name?—the one with the oversized shoes, starts the balloon animal gauntlet.” “Larry the Laugh‑inator,” Bill answered, tapping his watch. “He’ll be gone to the bathroom at 8:10, giving us a tiny window.” Molly nodded, her grin widening. “Fine. I’ll take the east side, you cover the west. We meet at the unicorn piñata—” “—and smash it for a quick exit,” Bill finished, chuckling. “If all goes well, we’ll be out with more loot than a kid’s birthday wish list. If it goes wrong, we’ll just blame the clown.” The two thieves exchanged a conspiratorial glance, then slipped into the party’s chaos like two mischievous elves, ready to pilfer joy and frosting in equal measure.
created_at
"2026-04-24T18:10:41.83Z"
4/24/2026, 6:10:41 PM
model
"openai/gpt-oss-120b"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
422
moderation_latency
(null)
generation_time
27921
tokens_prompt
77
tokens_completion
1250
native_tokens_prompt
114
native_tokens_completion
1048
native_tokens_completion_images
(null)
native_tokens_reasoning
317
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"
usage
0.0002677
router
(null)
provider_responses
0
endpoint_id
"85835306-dba3-4334-87bc-ab9ffa5c9c62"
is_byok
false
latency
86
model_permaslug
"openai/gpt-oss-120b"
provider_name
"DeepInfra"
status
429
1
endpoint_id
"3dc95f88-3960-4c91-b2bb-bbd2f5069c48"
id
"0500914b580729acd860ddb316c4d528"
is_byok
false
latency
354
model_permaslug
"openai/gpt-oss-120b"
provider_name
"Novita"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.1; linux; x64))"
http_referer
(null)
request_id
"req-1777054241-ZPvVj939LHIWRmv4ohjM"
session_id
(null)
api_type
"completions"
id
"gen-1777054241-86BuAITmVla1bHXcnvBU"
upstream_id
"0500914b580729acd860ddb316c4d528"
total_cost
0.0002677
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Novita"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
561
20.26%
Dialogue to Total Word Ratio
Ratio: 64.35%, Deviation: 5.65%
neededClean
false
wordsTotal
561
wordsDialogue
361
10.1304%