NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-01-01M1EPDPGXZYTHWRQQ8AAJ25NM
Run Details
Model:
Grok 4.6 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
2m 30s
Cost:
$0.046800
Provider:
xAI
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
87.30%
Dialogue to Total Word Ratio
(details)
43.6529%
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 old warehouse served as the perfect backdrop for Molly and Bill's latest scheme, its walls adorned with posters of failed heists that they now laughed about over cheap beer. Molly, a petite woman with a knack for turning ordinary objects into tools of the trade, stood before a large corkboard covered in photos of the target location: a modest house with a spacious backyard where little Emma's seventh birthday party was set to take place the following Saturday. Bill, her longtime partner whose bulk belied a surprisingly agile mind when it came to improvisation, sat nearby assembling a device that looked suspiciously like a modified party favor. They had chosen this job because it was unexpected, the kind of heist that would cement their reputation as the jesters of the criminal world. No high-security vaults or armed response teams, just a gathering of children, parents, and an abundance of wrapped gifts piled high on a picnic table. The comedy of it all appealed to Molly's sense of humor, imagining the look on the kids' faces when they realized the presents had vanished into thin air, or rather, into their van. "Alright, here's the plan," Molly announced, pointing to a diagram she had drawn. "We arrive early as the hired entertainment. I'll do balloon animals that look nothing like animals, you do card tricks that fail spectacularly. When the piƱata breaks, that's our cue. Chaos ensues, we fill these extra large gift bags with the real presents and swap them with empty boxes we brought." Bill looked up from his work, a screwdriver in hand. "Molly, swapping boxes? That's genius but what if a kid opens one early and finds nothing? We'll have a meltdown on our hands." "That's the fun part. We blame it on the magic. 'Abracadabra, the gifts went to the North Pole early!' Then we dash." They discussed the details at length, Molly explaining how they would use the bouncy castle as a hiding spot for the bags until the coast was clear. Bill suggested wearing disguises so elaborate that even their own mothers wouldn't recognize them, complete with fake noses and wigs in neon colors. The conversation turned to potential hitches. "What if it rains?" Bill asked. "Then everyone goes inside, harder to operate." "We pray for sun or bring umbrellas as props," Molly shot back. "And if a parent gets suspicious, we start a conga line to distract." Hours passed as they refined the strategy, testing gadgets like a remote-controlled toy car to fetch items from afar and practicing their routines until they were both in stitches from the failed attempts. Bill's balloon dog looked more like a deflated sausage, causing Molly to double over in laughter. "This is going to be epic," she said. "We'll be legends." "Or in jail eating cake with a bunch of toddlers," Bill retorted, but he was smiling. By dawn, the plan was set, a masterpiece of comedic crime that they hoped would go off without a hitch, or at least with enough hilarity to make the inevitable mishaps worthwhile. They packed their gear, the warehouse echoing with their banter, two thieves ready to steal the show and the presents.
created_at
"2026-09-01T14:36:56.74Z"
9/1/2026, 2:36:56 PM
model
"x-ai/grok-4.6-20260810"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
555
moderation_latency
(null)
generation_time
149559
tokens_prompt
75
tokens_completion
1618
native_tokens_prompt
267
native_tokens_completion
7775
native_tokens_completion_images
(null)
native_tokens_reasoning
7112
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
0.0468
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788273416-2I7pF0k6ga4TEtupdyu5"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788273416-4owiyGobjlNkU3Ovv7Dv"
upstream_id
"5f20deae-990f-9deb-afcc-ae40059d5f94"
provider_responses
0
endpoint_id
"0d0536e3-7eb1-4acb-8249-c8813365c2d8"
id
"5f20deae-990f-9deb-afcc-ae40059d5f94"
is_byok
false
latency
153
model_permaslug
"x-ai/grok-4.6-20260810"
provider_name
"xAI"
status
200
total_cost
0.0468
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
532
87.30%
Dialogue to Total Word Ratio
Ratio: 31.65%, Deviation: 1.65%
neededClean
false
wordsTotal
534
wordsDialogue
169
43.6529%