NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-10-13-01M4TBQ1VFA5RDDZ5PVSQ93W78
Run Details
Model:
Ember-1 (Reasoning, Max)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
2m 22s
Cost:
$0.135977
Provider:
Fireworks
Evaluation summary
Score
Evaluator
Details Link
86.38%
Matches word count
(details)
91.00%
Dialogue to Total Word Ratio
(details)
88.6907%
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 unrolled the blueprint across the hood of the Buick and weighed the corners down with a juice box, a sock, and two halves of a bagel. A clown wig drooped on the dashboard like a murdered parrot. "This is the Henderson residence. Saturday, two to four p.m. One birthday boy, fifteen guests, an estimated twenty-two presents." "Twenty-two," Bill whispered. "That's the big score." "Focus. Presents come in the front door, get stacked here. Gift table, by the cake." "What's the cake situation?" "Bill." "I'm just saying, if we're already there—" "We are not stealing cake. We are professionals." Bill nodded slowly, then raised his hand. He was already wearing the nose, which Molly had chosen not to address. "What if the cake is, like, right there?" "Here's the plan. You go in as a clown." "I don't know clown stuff." "You honk a horn and fall over. It's not Shakespeare." "And you?" "I'm the balloon animal artist. I case the room while I work." "Can you make balloon animals?" "I can make a snake." "That's just a balloon." "That's why it's convincing." "What if they ask for a dog?" "Make a snake and call it a dachshund." Bill leaned over the blueprint and squinted at a square drawn in crayon in the backyard. Three weeks of surveillance had gone into that crayon. "Bouncy castle?" "Escape route. Things go bad, we bounce." "You want to escape through the bouncy castle." "Name one cop who follows a guy into a bouncy castle." Bill opened his mouth, closed it, and nodded. "Now, security. Greg, the birthday dad: camcorder, hero complex. And Grandma Linda. Grandma Linda sees everything." "How do we handle Grandma Linda?" "Butterscotch. Put a bowl between her and the presents, she's out for forty minutes." "What about the kids?" "Kids notice things adults don't. A kid sees you stuffing a present down your clown pants, that kid tells everybody." "So what's the move?" "Piñata. We grab during the piñata. Every eye is on a donkey full of candy." "And the birthday boy?" Molly hesitated, which Bill had never once seen her do on a job. "He's seven, Molly." "Seven-year-olds have nothing to lose. Most dangerous people on earth." "Maybe we should rob a bank instead." "A bank has alarms and dye packs. A birthday party has twenty-two unguarded presents and a clown-shaped hole in its security that nobody has ever thought to exploit." Bill considered this, turning it over with the full resources of his brain. "What if one of the presents is socks?" "Then some kid dodged a bullet." "What if they're all socks? My Aunt Denise gave me socks every year till I was thirty." "Your Aunt Denise is why you're like this." Bill nodded, conceding the point. "When do we start clown training?" Molly checked her watch. "Now. Honk the horn. Fall over." Bill honked the horn and fell over. It was, she had to admit, not bad. "We're going to be rich," he said from the driveway. "We're going to be moderately rewarded. Keep your voice down."
created_at
"2026-10-13T10:07:47.576Z"
10/13/2026, 10:07:47 AM
model
"fireworks/ember-1-20260923"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
448
moderation_latency
(null)
generation_time
142002
tokens_prompt
75
tokens_completion
8153
native_tokens_prompt
149
native_tokens_completion
9062
native_tokens_completion_images
(null)
native_tokens_reasoning
8153
native_tokens_cached
148
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"
service_tier
(null)
usage
0.1359774
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1791886067-YCFbkvXzRmbgwlgAYsvN"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1791886067-hsPOUTUMpfWGUaZWGjAt"
upstream_id
"chatcmpl-2133d519e814436d8d470c90672142ac"
provider_responses
0
endpoint_id
"b89a2f05-4051-4446-a9b0-7bc8ae7b7c90"
id
"chatcmpl-2133d519e814436d8d470c90672142ac"
is_byok
false
latency
448
model_permaslug
"fireworks/ember-1-20260923"
provider_name
"Fireworks"
status
200
total_cost
0.1359774
cache_discount
0.0003996
upstream_inference_cost
0
provider_name
"Fireworks"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
86.38%
Matches word count
n/a
neededClean
false
words
511
91.00%
Dialogue to Total Word Ratio
Ratio: 71.37%, Deviation: 1.37%
neededClean
false
wordsTotal
517
wordsDialogue
369
88.6907%