NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-20-01KHYBAPTFB3PYARYSAFWHTZHG
Run Details
Model:
Grok 4 Fast
Test:
Dialogue tags
Scenario:
Write 200 words with 50% dialogue
Duration:
3.1s
Cost:
$0.000274
Provider:
xAI
Evaluation summary
Score
Evaluator
Details Link
95.99%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
47.9934%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 200 word scene that has 50% 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 cluttered basement, blueprints of the community center spread across a rickety table. Bill lounged on a beanbag, munching stolen donuts, crumbs dusting his fake mustache. "Listen up, Bill," Molly said, jabbing the map. "The party's at noon. Pinata, cake, squealing brats everywhere. We hit during the magician's act—distraction gold." Bill grinned, wiping his mouth. "Genius. I dress as the clown, you as a balloon animal. Sneak in, stuff presents in oversized shoes, waddle out." Molly snorted. "Oversized shoes? You'll trip over your own ego. No, we use the gift table. I'll decoy with a song about 'sharing is caring,' you grab the loot." "Sharing? Ha! Those kids won't share their snotty noses, let alone Teletubbies sets." Bill leaned forward. "What about the parents? Buzzkills with iPhones snapping pics." "Easy. Fake allergy attack—sneeze confetti everywhere. Chaos covers our exit." Molly sketched a getaway van labeled "Party Poopers Anonymous." Bill chuckled. "Poetic. But if a toddler clings to my leg?" "Kick it off gently. We're thieves, not monsters." She high-fived him. "This haul funds our Vegas spree. Operation Candy Heist is go." Bill saluted. "To sticky-fingered success!"
created_at
"2026-02-20T20:19:46.14Z"
2/20/2026, 8:19:46 PM
model
"x-ai/grok-4-fast"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1371
moderation_latency
(null)
generation_time
2800
tokens_prompt
77
tokens_completion
318
native_tokens_prompt
216
native_tokens_completion
507
native_tokens_completion_images
0
native_tokens_reasoning
215
native_tokens_cached
151
num_media_prompt
(null)
num_input_audio_prompt
(null)
num_media_completion
0
num_search_results
(null)
origin
"https://ncbench.com/"
is_byok
false
finish_reason
"stop"
native_finish_reason
"completed"
usage
0.00027405
router
(null)
provider_responses
0
endpoint_id
"c9586d9d-97ff-40ca-ae94-22504d6f9b7e"
id
"fed667c3-e9b1-1ab8-11df-ef609309c6a8"
is_byok
false
latency
465
model_permaslug
"x-ai/grok-4-fast"
provider_name
"xAI"
status
200
api_type
"completions"
id
"gen-1771618786-KL8QKByqvn5kjXPkNaC9"
upstream_id
"fed667c3-e9b1-1ab8-11df-ef609309c6a8"
total_cost
0.00027405
cache_discount
0.00002265
upstream_inference_cost
0
provider_name
"xAI"
Evaluation details
Result
Evaluator
Details
Meta Data
95.99%
Matches word count
n/a
neededClean
false
words
192
0.00%
Dialogue to Total Word Ratio
Ratio: 70.31%, Deviation: 20.31%
neededClean
false
wordsTotal
192
wordsDialogue
135
47.9934%