NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-01-01M1ENRC7SKRCRR82JPNJQPWT0
Run Details
Model:
Grok 4.6 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
2m 39s
Cost:
$0.041034
Provider:
xAI
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.13%
Dialogue to Total Word Ratio
(details)
0.0660%
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 paced the warehouse floor, excitement in her eyes, as Bill sat polishing his tools. "We're hitting the children's party tomorrow, Bill. Steal all the presents. It's the score of a lifetime, wrapped in ribbons and bows," Molly declared. "A lifetime of what? Jail time? Those parties have cameras now, doorbell ones," Bill retorted. "We'll wear masks under our costumes. Superhero masks, fitting the theme. No one will know it's us," she countered. "Superheroes stealing? Irony. Anyway, how do we carry them all? There could be fifty presents," he asked. "We bring large sacks, like Santa, but it's not Christmas. Call it party favors in reverse," Molly said. "Santa in July? Suspicious. Better as the cleanup crew after the party, but that's too late, presents opened," Bill noted. "No, during. We volunteer to help set up, then help ourselves," she planned. "Volunteer? They'll have our names. Fake IDs needed," he pointed out. "Already got them. I'm Mary, you're Bob. Simple names," Molly replied. "Mary and Bob, the present snatchers. Catchy. Now, the break in part. Is it break in or walk in?" Bill questioned. "Walk in, invited by our fake connection. I know a guy who knows the mom," she explained. "A guy? Reliable? Last guy ratted us out," he reminded. "This one's solid. Anyway, once in, locate the present pile, usually in the living room. Guard it with our bodies, then when game time, move them to the kitchen, then out the back," Molly detailed. "Kitchen to back door. Is there a back door? Yes, from casing," Bill confirmed. "Yes. We did the recon last week. Fence is low, hop it with sacks," she said. "Sacks over fence, like burglars. Classic. But kids might see and think it's a game, follow us," he worried. "Then we say it's a treasure hunt, the presents are the treasure, but we keep them," Molly laughed. "Too risky. Better a smoke screen, but not real smoke, like a fog machine for the party that we 'fix' by making fog, hide in it," Bill suggested. "Fog machine? Do they have one? We bring it as extra entertainment. 'Surprise fog for the photos!' Then grab in the mist," she agreed. "Genius. Comedic fog heist. I like it. What if it sets off alarms or sprinklers?" he asked. "No, it's a house party, not that advanced. We'll be fine. Practice the fog tomorrow morning," Molly instructed. "And the presents, we unwrap later, see what's good. Hope for cash or gift cards too," Bill hoped. "Yes, mixed in. The invitation said no toys over 20 dollars, but parents always overspend," she noted. "Good. Split even, as always. This could fund our next vacation heist, like a cruise or something," he dreamed. "First this one. Focus. We go over the timeline: arrive 1:45, set up 2:00, fog at 2:30, grab 2:35, out by 2:40," Molly timed. "Tight schedule. What if the birthday kid is late?" Bill asked. "Improvise. We're professionals, comedic professionals," she said. "Alright, I'm in. Let's steal those presents and make a funny story out of it," Bill concluded with a smile. The thieves spent hours more discussing contingencies like rain or extra guests, their dialogue filled with puns and jokes about wrapping the crime up neatly.
created_at
"2026-09-01T14:25:18.083Z"
9/1/2026, 2:25:18 PM
model
"x-ai/grok-4.6-20260810"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
513
moderation_latency
(null)
generation_time
158466
tokens_prompt
75
tokens_completion
2306
native_tokens_prompt
267
native_tokens_completion
6814
native_tokens_completion_images
(null)
native_tokens_reasoning
6046
native_tokens_cached
256
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
"default"
usage
0.041034
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788272718-j6lvwRfHDAc9shsQjV0I"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788272718-HXUjnFhdNaAeLFvUSYnL"
upstream_id
"b22bb83f-dcee-9ff4-8bc4-76eec0aaec9c"
provider_responses
0
endpoint_id
"0d0536e3-7eb1-4acb-8249-c8813365c2d8"
id
"b22bb83f-dcee-9ff4-8bc4-76eec0aaec9c"
is_byok
false
latency
152
model_permaslug
"x-ai/grok-4.6-20260810"
provider_name
"xAI"
status
200
total_cost
0.041034
cache_discount
0.000384
upstream_inference_cost
0
provider_name
"xAI"
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
541
0.13%
Dialogue to Total Word Ratio
Ratio: 81.52%, Deviation: 11.52%
neededClean
false
wordsTotal
541
wordsDialogue
441
0.0660%