NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-02-26-01KJCRYNHCB1E743MGDJXBQGVZ
Run Details
Model:
LFM2 24B
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
16.2s
Cost:
$0.000104
Provider:
Together
Evaluation summary
Score
Evaluator
Details Link
35.00%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
17.5011%
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
The moon hung low over Maplewood Park, casting silver light on the inflatable unicorn bouncing near the party tents. Molly crouched behind a hedge, her eyes darting between the security camera and the glowing red door of the children’s birthday bash. Beside her, Bill adjusted his ski mask, trying to look menacing but failing to hide his nervous twitch. “Alright,” Molly whispered, “the plan is simple. We slip in before the piñata gets smashed, grab every present within thirty seconds, and vanish like ghosts.” She pulled out a crumpled blueprint marked with arrows and question marks. Bill snorted. “Ghosts? We’re not even wearing proper masks. The kids’ll spot us faster than a cat chasing a laser pointer.” Molly ignored him, flipping through her notes. “The host, Mr. Higgins, leaves for a bathroom break at exactly eight-fifteen. That’s our window. But we need to disable the motion sensors first.” “Sensors?” Bill raised an eyebrow. “You mean the ones that blink red when someone walks by? I thought those were just for show.” “Exactly,” Molly said, grinning. “They’re wired to trigger the alarm if anyone lingers too long. We’ll need a distraction.” Bill leaned closer, lowering his voice. “Distraction? Like… throwing confetti?” “Better,” Molly replied. “We could ‘accidentally’ knock over the cake table. The mess will draw everyone’s attention while we slip into the gift room.” Bill laughed nervously. “You’re insane. What if the parents see us?” “Then we blame the dog,” Molly said, pointing to a shaggy golden retriever tied to a post nearby. “Mr. Wuffles here loves chaos.” Bill shook his head. “That dog’s more likely to chase its tail than help us.” “Exactly!” Molly clapped her hands. “Now, about the presents—there’s a mountain of them. Stuffed animals, puzzles, glow-in-the-dark slime. All the good stuff.” Bill squinted. “You sure this isn’t overkill? We’re stealing toys from kids’ parties. What if we get caught?” “Caught?” Molly scoffed. “We’re professionals. Besides, the real risk is getting stuck in the bounce house if we trip the sensor.” Bill groaned. “You’re impossible. What if someone asks where we’re going?” “Simple,” Molly said, pulling out a fake mustache from her pocket. “We tell them we’re there for the ‘special guest’—a clown named Clive. He’s… uh… very shy.” Bill rolled his eyes. “Clive the Clown? That’s not a thing.” “Exactly! No one expects a clown. They’ll be too busy laughing at my mustache to notice us.” A distant shout echoed from the party area. “Is that… a balloon popping?” Molly froze. “That’s the signal. Let’s move.” Bill grabbed the backpack containing their tools—a set of lockpicks, a flashlight, and a bag of candy corn. “Wait, why do we have candy corn? This isn’t a heist, it’s a snack run.” “Strategic loot,” Molly said, tossing him a stuffed unicorn. “Now, remember: quiet as mice, fast as lightning. And if anything goes wrong, we blame Mr. Higgins’ parrot.” Bill saluted half-heartedly. “Parrot’s got better aim than me.” They crept toward the party tent, hearts pounding. Somewhere inside, a child giggled, unaware their gifts were about to become part of a very odd heist story.
created_at
"2026-02-26T10:47:13.54Z"
2/26/2026, 10:47:13 AM
model
"liquid/lfm-2-24b-a2b-20260224"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
334
moderation_latency
(null)
generation_time
16157
tokens_prompt
77
tokens_completion
915
native_tokens_prompt
75
native_tokens_completion
845
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
0.00010365
router
(null)
provider_responses
0
endpoint_id
"b5f259c3-1e77-40d6-8fa4-161cdf0e4a7c"
id
"oYufEJj-2j9zxn-9d3edcf5b614e508"
is_byok
false
latency
334
model_permaslug
"liquid/lfm-2-24b-a2b-20260224"
provider_name
"Together"
status
200
api_type
"completions"
id
"gen-1772102833-MIG0ARCEaW08km0j53Jl"
upstream_id
"oYufEJj-2j9zxn-9d3edcf5b614e508"
total_cost
0.00010365
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Together"
Evaluation details
Result
Evaluator
Details
Meta Data
35.00%
Matches word count
n/a
neededClean
false
words
518
0.00%
Dialogue to Total Word Ratio
Ratio: 60.23%, Deviation: 30.23%
neededClean
false
wordsTotal
523
wordsDialogue
315
17.5011%