NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1K4T59XAWXDQXHJVB9ZVADB
Run Details
Model:
Muse Spark 1.3 (Reasoning, Medium)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
2m 4s
Cost:
$0.029572
Provider:
Meta
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
95.04%
Dialogue to Total Word Ratio
(details)
47.5177%
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 crouched behind the minivan and peered through the fence at the sagging bouncy castle in the backyard. Bill adjusted his squeaky clown nose and tried to look inconspicuous behind a trash can. Molly whispered, "Okay, visual confirmed. That is a mountain of presents by the cake table." Bill hissed, "I see twenty-three. One is definitely bike-sized. God bless seven-year-olds with rich grandparents." A chorus of screaming kids chased a dog wearing a party hat around the yard. Molly muttered, "Focus. The plan is simple. We walk in as entertainment, grab the loot during pinata chaos, and roll out." Bill grinned, "I still think my clown cover is flawless. Kids love clowns. Adults fear clowns. Perfect distraction." Molly yanked the rainbow wig off his head and stuffed it into the duffel bag. "No wigs. Last time you scared a toddler and he bit you. We are going as tired parents. Much more invisible." "Do I look like a tired parent? I feel like I radiate single man who owns ferrets." "Put applesauce on your shirt, slouch, and complain about nap schedules. You will blend right in." Bill smeared a juice box on his shirt and practiced his exhausted sigh. "Like this? Ugh, Brayden did not sleep last night and Paw Patrol is too loud?" "Perfect. You sound miserable. Now once we are inside, you handle the moms by the gift table." "How do I handle them? I cannot just say nice wrapping paper and shove it in a bag." "You talk about school districts and nut allergies. They will form a circle and trap you for an hour." The clown music stopped and someone shouted that the magician had cancelled. "Oh, that is our cue. We can be the backup magicians. I know one card trick and a lot of confidence." "No magic. Magic invites attention. We want the opposite of attention. We want beige energy." "Fine. Operation Beige Dad. What about the presents? They are guarded by a grandma with eagle eyes." Molly pulled out two empty IKEA bags and a birthday card addressed to someone named Jayden. "Grandmas love helpers. I will say Jayden is my nephew and we forgot to add our gift, so we need to rearrange the pile." "Then while she tells you about her hip replacement, I slide the bags under the table like a present ninja?" "Exactly. When the pinata breaks, every kid screams, every parent films, and we walk out with twenty-three gifts." A balloon popped and confetti drifted over the fence onto their getaway stroller. "Is stealing from a seven-year-old a new low? Even for us? We once robbed a mime convention." "We are not stealing joy, just boxes. Plus his dad drives a Tesla. He will recover emotionally by Christmas." "Okay, but if there is a bouncy castle, I get one bounce. For morale. For stealth testing." "One bounce. No shoes. And if you pop it, you are buying twenty-three replacement presents yourself." Bill stood up, brushed chips off his beige dad disguise, and grabbed the stroller handle like a pro. "Let's go steal some birthday. Operation Beige Dad is a go." They waddled toward the gate waving the card while kids shrieked and frosting flew everywhere.
created_at
"2026-09-03T08:05:22.885Z"
9/3/2026, 8:05:22 AM
model
"meta/muse-spark-1.3-20260902"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3726
moderation_latency
(null)
generation_time
124468
tokens_prompt
75
tokens_completion
956
native_tokens_prompt
68
native_tokens_completion
6938
native_tokens_completion_images
0
native_tokens_reasoning
6059
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
"completed"
service_tier
"auto"
usage
0.0295715
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788422722-JIZfiye98wi98PlK4NYn"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788422722-f1DT8H2n49pRFmMX6Ymj"
upstream_id
"resp_6a992a43e614d1be6f8146d2"
provider_responses
0
endpoint_id
"cf4f1b4e-1719-4b65-9111-7dd7635e5a2f"
id
"resp_6a992a43e614d1be6f8146d2"
is_byok
false
latency
624
model_permaslug
"meta/muse-spark-1.3-20260902"
provider_name
"Meta"
status
200
total_cost
0.0295715
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Meta"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
537
95.04%
Dialogue to Total Word Ratio
Ratio: 71.01%, Deviation: 1.01%
neededClean
false
wordsTotal
545
wordsDialogue
387
47.5177%