NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-02-26-01KJCRZ40W3Z4AFZ3GGRAVKRJ7
Run Details
Model:
LFM2 24B
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
15.7s
Cost:
$0.000099
Provider:
Together
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
2.16%
Dialogue to Total Word Ratio
(details)
1.0799%
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 living room buzzed with the muffled giggles of kids at the birthday bash next door. Molly leaned against the doorway, arms crossed, eyes darting between the glittery decorations and the clock ticking toward midnight. Bill crouched beside her, clutching a duffel bag stuffed with lockpicks and a stolen party hat. “Alright,” Molly said, tapping her chin, “we hit at 11:30. The parents leave early, right? That’s our window.” Bill nodded, adjusting his glasses. “Easy in, easy out. The back gate’s unguarded till then. But we need a distraction.” He pulled a crumpled note from his pocket. “This is the new trick—my cousin’s kid’s party. She’s got a pet iguana named Steve. We’ll ‘rescue’ Steve, and everyone’ll rush outside.” Molly snorted. “A iguana? You’re stealing presents for a lizard?” “Steve’s a *star*, Molly. Plus, kids love animals. They’ll be too busy cooing over him to notice the real loot.” She rolled her eyes but smirked. “Fine. But how do we get past the ‘Toy Terror’ banner?” Bill grinned. “Simple. I’ve got a magnet for the metallic parts of the display. And you’ll distract the mom with your ‘cute’ act.” “My ‘cute’ act?” “You know—the one where you pretend to cry over a broken toy. It’s classic.” Molly fake-gasped. “You think I’d cry over a plastic dinosaur?” “You’d sob like a lost puppy if it meant getting to the presents.” She sighed, then straightened. “Okay, but what if they check the gift table first?” Bill held up the duffel. “We swap the gifts. Take the ones under the table, leave the rest. No one’ll notice unless they’re looking.” “And if someone sees us?” “They’ll think it’s part of the party. Kids love chaos. Besides, we’re ghosts.” Molly laughed. “Ghosts with lockpicks.” “Exactly.” He paused. “Wait, what about the cake?” “Don’t even think about eating it. We’re thieves, not vandals.” “Not yet,” Bill said, pulling out a tiny flashlight. “But if we need to, we’ll make excuses. ‘Oh, this candle’s lit!’ ‘Look, a mouse!’” Molly mock-scowled. “You’re ridiculous.” “Ridiculously effective.” He tossed her a balloon animal shaped like a shark. “Here. For morale.” She caught it, then frowned. “Why a shark?” “Because sharks don’t judge. They just… bite.” They both cracked up, the sound echoing through the hallway. Molly pocketed the shark, then straightened. “Alright, plan’s set. Meet me at the gate in ten. And Bill?” “Yeah?” “Don’t let Steve escape.” He saluted. “No promises, Captain.” Outside, the night air bit as they crept toward the backyard. Molly glanced at the party lights, then whispered, “What if the iguana steals *our* presents?” Bill smirked. “Then we’ll steal Steve’s. He’s got better taste anyway.” They vanished into the shadows, the duffel bag bouncing between them, ready to turn a children’s party into their personal treasure hunt.
created_at
"2026-02-26T10:47:28.376Z"
2/26/2026, 10:47:28 AM
model
"liquid/lfm-2-24b-a2b-20260224"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
242
moderation_latency
(null)
generation_time
15554
tokens_prompt
77
tokens_completion
851
native_tokens_prompt
75
native_tokens_completion
809
native_tokens_completion_images
(null)
native_tokens_reasoning
0
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
0.00009933
router
(null)
provider_responses
0
endpoint_id
"b5f259c3-1e77-40d6-8fa4-161cdf0e4a7c"
id
"oYufJia-2j9zxn-9d3edd5272378131"
is_byok
false
latency
242
model_permaslug
"liquid/lfm-2-24b-a2b-20260224"
provider_name
"Together"
status
200
api_type
"completions"
id
"gen-1772102848-CFGAlxkg8qxCkVJrmKYu"
upstream_id
"oYufJia-2j9zxn-9d3edd5272378131"
total_cost
0.00009933
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Together"
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
465
2.16%
Dialogue to Total Word Ratio
Ratio: 61.24%, Deviation: 8.76%
neededClean
false
wordsTotal
467
wordsDialogue
286
1.0799%