NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-04-24-01KQ079ARP1FWW5T6R4VZYVSBV
Run Details
Model:
GPT-OSS 120B
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
1m 10s
Cost:
$0.002968
Provider:
Together
Evaluation summary
Score
Evaluator
Details Link
75.16%
Matches word count
(details)
3.03%
Dialogue to Total Word Ratio
(details)
39.0927%
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
Molly and Bill crouched behind a hedge, flashlights flickering like nervous fireflies. The cul‑de‑sac was quiet, save a distant bark and a soft lullaby drifting from a house. Molly adjusted her ridiculous plastic spy cap, while Bill tightened the straps on his oversized backpack meant for schoolbooks, not loot. Balloon animals and glittering gift wrap towered around them under the streetlamp. A sign read “Happy Birthday, Timmy! (Age 5)”, and chocolate cake scent floated like temptation. Both tried to keep straight faces, but the absurdity made their lips twitch upward in a conspiratorial grin. Molly: "Alright, Bill, the plan is simple—sneak in through the puppy’s tunnel, grab the presents, and vanish before the clown notices." Bill: "Did you say puppy’s tunnel? Are we expecting a Labrador with a secret passage or just a cardboard box?" Molly: "It’s a literal cardboard tunnel shaped like a puppy, left over from last year’s “Puppy Party” theme—perfect camouflage." Bill: "Perfect, because nothing says ‘stealth’ like waddling through a giant, squeaky dog that smells like cheese and crayons." Molly: "Just remember, the kids will be distracted by the balloon pop contest, so we have a five‑minute window." Bill: "Five minutes? That’s about the time it takes for a toddler to discover glitter on the floor and start a panic." Molly: "Exactly why we need the glitter‑neutralizing spray I stole from the craft store—no sparkle, no alarm." Bill: "And the “present‑grabber” gloves—those oversized mitts that make everything look like a giant, clumsy snowball." Molly: "Don’t forget the decoy: the inflatable unicorn that will inflate itself if anyone approaches, buying us extra seconds." Bill: "An inflatable unicorn? That’s the most majestic guard dog I’ve ever imagined, and it’s also ridiculous." Molly: "Now, what about the cake? We can’t leave it untouched—no thief respects a frosting‑filled masterpiece." Bill: "I propose we eat it. Nothing says ‘no witnesses’ like a sugar high and sticky fingers." Molly: "Deal, but after the cake we’ll need a quick exit—maybe the laundry chute the parents never use." Bill: "A laundry chute? That’s the classic thief’s route—right into a basket of socks and regret." Molly: "Regret is optional if we time it with the fire alarm drill—chaos guarantees a clean getaway." Bill: "Fire alarm drill? You mean the one the homeowners scheduled for exactly 7:45? Talk about convenient." Molly: "Exactly, the moment the kids start screaming, the adults will be too busy hiding, and we’ll be king of the loot." Bill: "King of presents? I always imagined a throne made of crayons and bubble wrap, not stolen toys." Molly: "Your imagination is the only thing stopping us from being legendary—so let’s get legendary, partner." Bill: "To legendary heists and stolen birthdays—let’s make this the most unforgettable party nobody will remember." Molly: "One last thing—if the kidnapper of the cupcakes shows up, we’ll negotiate with cookies, not surrender. And we’ll hand them the empty gift boxes as evidence of our generosity." Bill: "Deal. And if we get caught, we’ll blame the inflatable unicorn; it’s the obvious suspect."
created_at
"2026-04-24T17:06:04.703Z"
4/24/2026, 5:06:04 PM
model
"openai/gpt-oss-120b"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
213
moderation_latency
(null)
generation_time
69558
tokens_prompt
77
tokens_completion
4930
native_tokens_prompt
127
native_tokens_completion
4915
native_tokens_completion_images
(null)
native_tokens_reasoning
4075
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
0.00296805
router
(null)
provider_responses
0
endpoint_id
"c7cc0654-cfe3-4531-bd8a-68cd162ba020"
id
"ogSrcx3-z1gNr-9f16b24b8796b208"
is_byok
false
latency
190
model_permaslug
"openai/gpt-oss-120b"
provider_name
"Together"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.1; linux; x64))"
http_referer
(null)
request_id
"req-1777050364-p9nDUmiyPPjJrOQykTxP"
session_id
(null)
api_type
"completions"
id
"gen-1777050364-kxYxwAHhxIpDe3tnqoOH"
upstream_id
"ogSrcx3-z1gNr-9f16b24b8796b208"
total_cost
0.00296805
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Together"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
75.16%
Matches word count
n/a
neededClean
false
words
513
3.03%
Dialogue to Total Word Ratio
Ratio: 78.36%, Deviation: 8.36%
neededClean
false
wordsTotal
513
wordsDialogue
402
39.0927%