NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-04-24-01KQ03A75H6BD3SJEHKJAQPH7B
Run Details
Model:
GPT-OSS 120B
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
40.3s
Cost:
$0.000482
Provider:
SiliconFlow
Evaluation summary
Score
Evaluator
Details Link
75.16%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
37.5778%
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 pressed her nose against the grimy window of the van, squinting at the brightly colored house across the street. The sun was setting, casting a lazy orange glow over the cul‑de‑sac where the neighborhood kids’ birthday party was in full swing. Balloons bobbed like over‑inflated jellyfish, and a squeaky version of “Happy Birthday” drifted from the open patio door. Inside, the scent of melted chocolate and balloon‑rubber mingled with the faint perfume of candy floss, turning the air into something that could only be described as a sugary nightmare for a seasoned burglar. Bill slumped into the driver’s seat, feet propped on the dashboard, his cap tilted back so the Wi‑Fi‑enabled earpiece was visible. “You see the cake? It’s a triple‑layer chocolate disaster. That’s our first target.” He tapped the steering wheel with a thumb, mimicking a drumroll. Molly laughed, a nervous giggle that sounded like a hiccup. “We’re not here for the cake, Bill. We’re after the presents. The kids’ parents think they’re safe behind glitter‑covered wrapping paper. They’re not.” Bill glanced at the backyard gate, a flimsy wooden thing that looked like it hadn’t seen a lock since the 1990s. “I’ve got a plan. We’ll slip in through the dog’s kennel. The mutt’s name is Biscuit, right?” Molly nodded, tapping a finger against her cheek. “Sure. Biscuit won’t notice us if we’re quiet. He’s already busy chasing his tail in the garden. We just need a distraction for the adults.” Bill opened his notebook, the pages already stained with coffee rings. “I’m thinking of using the karaoke machine. We’ll crank up ‘Let It Go’—the kids will love it, the parents will get annoyed, and the chaos will cover our entrance.” Molly’s eyes sparkled. “And once we’re in, we’ll grab every wrapped box, every gift bag, and the whole thing—” Bill raised a hand. “—and we’ll leave a note that says ‘Thanks for the loot!’ in glitter. That’ll be the kicker.” Molly smirked, pulling a small, plastic keychain from her pocket. “I already stole the key to the party tent. It’s hidden under the flower pot by the fence. Easy access.” Bill checked his watch, the seconds ticking like a metronome. “We have exactly twenty‑nine minutes before the magician shows up. Timing is everything.” Molly paced, her shoes making soft scuffs on the van’s floor. “What about the security camera? The one that watches the entrance?” Bill shrugged. “We’ll do the classic ‘screen‑savvy’ move—plug the electric cord into a lamp and then pull the plug. The system will think it’s a power surge and reboot. Thirty seconds of blind spot.” Molly chuckled. “You’re a genius, Bill. We’ll be in and out before the kids can say ‘party!’” Bill grinned, tapping the steering wheel again. “Exactly. Now, let’s make this the most spectacular children’s‑party heist ever—” Molly leaned forward, eyes bright. “—and the only time the kids will ever say ‘thank you’ to a thief.”
created_at
"2026-04-24T15:56:39.485Z"
4/24/2026, 3:56:39 PM
model
"openai/gpt-oss-120b"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1888
moderation_latency
(null)
generation_time
40064
tokens_prompt
77
tokens_completion
1251
native_tokens_prompt
127
native_tokens_completion
1056
native_tokens_completion_images
(null)
native_tokens_reasoning
356
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
0.00048155
router
(null)
provider_responses
0
endpoint_id
"ce118593-d9b2-4d2d-bcd4-33f0609b42f1"
id
"019dc0352b85577a0b5c1a9b7fdd3f93"
is_byok
false
latency
1813
model_permaslug
"openai/gpt-oss-120b"
provider_name
"SiliconFlow"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.1; linux; x64))"
http_referer
(null)
request_id
"req-1777046199-BALzhjLEbuLpEjp0mlJG"
session_id
(null)
api_type
"completions"
id
"gen-1777046199-VASBjhTMdrqRIZvfoQuL"
upstream_id
"019dc0352b85577a0b5c1a9b7fdd3f93"
total_cost
0.00048155
cache_discount
(null)
upstream_inference_cost
0
provider_name
"SiliconFlow"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
75.16%
Matches word count
n/a
neededClean
false
words
487
0.00%
Dialogue to Total Word Ratio
Ratio: 54.62%, Deviation: 24.62%
neededClean
false
wordsTotal
487
wordsDialogue
266
37.5778%