NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-01-01M1EN53JBT5WY6MDS4NDR6EM8
Run Details
Model:
Grok 4.6 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
6m 50s
Cost:
$0.126252
Provider:
xAI
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
0.0001%
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 excitedly. Bill watched from his chair. "Bill, this children's party is the opportunity we've been waiting for. All those presents, just sitting there for the taking. We are going to break in and steal them all. My plan is we will go as the hired clowns. I have the costumes ready. We juggle, we make balloon animals, and while the kids are watching the balloons, we slip the presents into our oversized clown pants which have secret pockets. Then we honk our way out. It's perfect because clowns are expected, no one will question two extra clowns. The comedy writes itself when the parents find the presents gone and the clowns vanished." Bill shook his head. "Clowns? Those pants will be obvious if they're full of boxes. We'll look lumpy. I have an idea that's better. We wait until the piñata is brought out. I volunteer to hang it, but instead I hang a bag, and when they break it, instead of candy, it's a smoke bomb I planted, smoke everywhere, confusion, we take the presents in the smoke screen. Classic heist movie stuff, but with a piñata twist, so funny." Molly clapped her hands. "Smoke bomb in piñata! That's brilliant and hilarious. The kids expecting candy, getting smoke, us getting gifts. But we need to make sure the smoke is safe, not too much. We can buy the party smoke, the kind for effects. Then, to make it even better, we can have a second piñata with the real candy to not be too mean. No, that's extra. Anyway, we do the smoke, grab, go. If the smoke doesn't work, we have the clown backup. Two plans, good. The party is in the afternoon, we arrive with the 'entertainment' van, which is our getaway. Load the presents in the back while 'setting up'. Even before the party starts, some stealing. Then the rest during. This is going to be our best comedic heist. People will read about it and laugh. Two thieves using party tricks to trick the party. Genius." Bill smiled. "Okay, clowns and smoke piñata. Let's get the bomb and the pants. If we get caught, at least we'll look funny in court, two clowns stealing toys. Worth it for the story. Let's do this." They both started listing the supplies needed, talking over each other. Molly said, "We need red noses, big shoes, the smoke canisters that are legal for parties, the van cleaned out, extra bags, maybe some real balloons to make it look authentic. Also, a playlist of clown music to play as we leave, to add to the comedy. If we pull this off, we'll have enough to live on for months from selling the toys. Imagine the looks on those kids' faces, but they'll get over it, new toys next week. We are providing a lesson in not leaving presents unattended. That's our justification." Bill said, "Yes, and I need to practice juggling so I don't drop the 'balls' which are actually small presents. This is going to be a blast, literally with the smoke. Let's start now." They continued the conversation, filling in small details like what to do if a child asks for a specific balloon animal while they are stuffing pants, deciding to say the balloon ran away, adding to the humor of the situation.
created_at
"2026-09-01T14:14:46.613Z"
9/1/2026, 2:14:46 PM
model
"x-ai/grok-4.6-20260810"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
505
moderation_latency
(null)
generation_time
410151
tokens_prompt
75
tokens_completion
4550
native_tokens_prompt
267
native_tokens_completion
21017
native_tokens_completion_images
(null)
native_tokens_reasoning
20311
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
0.126252
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788272086-yfcPacSfG0KMgC0m482h"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788272086-9LwTWCka6TBo5NgoH7k9"
upstream_id
"30662fdc-6282-9ecc-811b-7a9e2d91d107"
provider_responses
0
endpoint_id
"0d0536e3-7eb1-4acb-8249-c8813365c2d8"
id
"30662fdc-6282-9ecc-811b-7a9e2d91d107"
is_byok
false
latency
143
model_permaslug
"x-ai/grok-4.6-20260810"
provider_name
"xAI"
status
200
total_cost
0.126252
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
561
0.00%
Dialogue to Total Word Ratio
Ratio: 86.45%, Deviation: 16.45%
neededClean
false
wordsTotal
561
wordsDialogue
485
0.0001%