NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-01-01M1EP7KVX0T3VBWX1SCSG7XDM
Run Details
Model:
Grok 4.6 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
1m 43s
Cost:
$0.024984
Provider:
xAI
Evaluation summary
Score
Evaluator
Details Link
68.10%
Matches word count
(details)
0.02%
Dialogue to Total Word Ratio
(details)
34.0601%
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 cluttered hideout, eyes on the house sketch. Bill sat chewing a sandwich, doubtful. "This kids' party is our big score, Bill. All those presents piled up, easy grab while everyone's distracted by cake and games," Molly said. "Stealing from children, Molly? That's a new low even for us. I thought we hit banks, not birthday bashes," Bill replied. "Banks have lasers and guards. This is balloons and juice boxes. We pose as the hired clowns. You juggle badly, I twist balloon animals that pop immediately," she explained. "I can't juggle. Last attempt I hit myself in the face. And popping balloons might scare the little ones," he worried. "That's the diversion. Chaos is our cover. While they scream at the pops, we fill these oversized party bags with wrapped gifts from the table," Molly insisted. "Oversized bags at a party look suspicious. Parents will ask why the clowns are hoarding loot," Bill pointed out. "We claim it's for a piƱata refill or leftover collection. Improvise like always. You trip over the punch bowl for extra mess," she suggested. "I always get the clumsy part. Why not just sneak in after dark when kids are gone?" he asked. "Presents leave with the guests. We strike mid-party, grab bulk, vanish before opening time. The Wilsons are loaded, gifts worth pawn shop gold," Molly argued. "Loaded or not, getting spotted by a toddler in a superhero shirt could blow it. Costumes help, but fairy wings on me? I'll look insane," Bill complained. "Insane is perfect. No one suspects the dancing fairy of theft. We enter the side gate like invited entertainers, set up near gifts," she said. "Near gifts means eyes on us. What if a parent wants a balloon giraffe right as I snatch a box?" he questioned. "Hand them a deflated one and dash. Timing with Happy Birthday singing, all faces to the cake. You start the conga to pull crowd away," Molly planned. "Conga line? I'm not leading kids in a dance while you bag toys. That's too much spotlight," Bill protested. "Then fake a magic trick fail, rabbits everywhere, fake of course. Confusion reigns, we exit kitchen door to the van two streets over," she countered. "Van loaded with stolen teddy bears and video games. Hope no unique items trace back. And if a cop dad is there?" he cautioned. "Unlikely at a five-year-old bash. We wear wigs under clown paint, change in the van. Profits fund the real museum job next," Molly beamed. "Museum after this farce. Fine, I'm in, but if we end up in a bouncy castle chase, your fault," Bill sighed. "Deal. Practice your balloon twist. Make it look accidental explosion. Legendary heist or hilarious fail, either way we laugh later," she concluded. They high-fived over pizza crusts, refining the clown waddle and bag stash spots, joking about possible cake-in-face getaways and dinosaur costume backups if clowns flopped.
created_at
"2026-09-01T14:33:37.413Z"
9/1/2026, 2:33:37 PM
model
"x-ai/grok-4.6-20260810"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
536
moderation_latency
(null)
generation_time
103269
tokens_prompt
75
tokens_completion
1219
native_tokens_prompt
267
native_tokens_completion
4139
native_tokens_completion_images
(null)
native_tokens_reasoning
3481
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
0.024984
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788273217-68uTbmg5f5yyPKnxqxk7"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788273217-2nQZ0OpwfdQ8UDl0H5WY"
upstream_id
"ea84aabf-4a5e-980c-9230-bc29caa03cf6"
provider_responses
0
endpoint_id
"67ed3d9b-4f05-4e79-9bb6-1e8423a6c276"
id
"ea84aabf-4a5e-980c-9230-bc29caa03cf6"
is_byok
false
latency
154
model_permaslug
"x-ai/grok-4.6-20260810"
provider_name
"xAI"
status
200
total_cost
0.024984
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
68.10%
Matches word count
n/a
neededClean
false
words
486
0.02%
Dialogue to Total Word Ratio
Ratio: 83.13%, Deviation: 13.13%
neededClean
false
wordsTotal
492
wordsDialogue
409
34.0601%