NC Bench
Overview
About
Tests
Models
Model
Claude Opus 4.6
Test
Dialogue tags
Scenario
Write 500 words with 30% dialogue
Run ID
2026-02-08-01KGYYJ0B9TX1QA36X5TP8SG89
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
Molly spread the hand-drawn map across the sticky surface of the diner table, pushing aside two plates of half-eaten waffles and a coffee cup that had seen better decades. The map featured an alarming number of crayon annotations, a juice stain in the northeast quadrant, and what appeared to be a dinosaur drawn in the margin. She had stolen the layout from the birthday boy's refrigerator during a reconnaissance mission disguised as a door-to-door cookie saleswoman. Bill leaned forward and squinted at the document with the intensity of a man studying the blueprints to Fort Knox, which was ironic given that the target was a suburban split-level in a cul-de-sac where the most dangerous security feature was a golden retriever named Biscuit. "So the presents will be here," Molly said, tapping a rectangle she had labeled GIFT TABLE in aggressive red ink. "Stacked and unguarded during the cake phase. That's our window." "How long is the cake phase?" Bill asked, pulling a pen from behind his ear and beginning to take notes on a napkin. "Depends on the candle situation. The kid is turning seven, so we've got seven candles plus the potential for a novelty candle shaped like a race car. I'm estimating ninety seconds from the singing of Happy Birthday to the first slice being served. Every child in that room will be laser-focused on frosting." Bill whistled low and slow. The timeline was tight, but he had pulled off worse. Last Easter he had stolen an entire basket raffle from a church fundraiser in under forty-five seconds, though he still carried the emotional scars from a nun who had nearly caught him in the parking lot. "What about the parents?" he asked. "Neutralized. There's a mimosa bar." Bill nodded with genuine respect. Molly always thought three moves ahead. She had once orchestrated the theft of every prize from a carnival ring toss by creating a diversion involving a rented pony and a strategically placed air horn. The woman was an artist. Molly pulled out two matching shirts from her bag and tossed one to Bill. They were purple with a cartoon magician on the front and the words PRESTO PARTY ENTERTAINMENT embroidered on the back. Bill held the shirt up and frowned at the magician's unsettling grin. "We go in as the entertainment," Molly explained. "I do card tricks. You handle balloon animals. While I'm performing for the kids in the living room, you excuse yourself, load the presents into the van through the side door, and we're gone before anyone realizes the magic show isn't on the original itinerary." "Do I actually have to make balloon animals?" "Can you?" "I can make a snake." "That's just an uninflated balloon, Bill." "It's a creative interpretation." Molly rubbed her temples. Every heist had its complications. She folded the map, tucked it into her jacket, and dropped a twenty on the table. "Learn the giraffe by Saturday," she said. "The whole operation depends on it."
created_at
"2026-02-08T15:40:26.090477+00:00"
Sunday, February 8th, 2026 at 3:40:26 PM GMT+00:00
model
"anthropic/claude-4.6-opus-20260205"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1381
moderation_latency
218
generation_time
18745
tokens_prompt
61
tokens_completion
614
native_tokens_prompt
75
native_tokens_completion
707
native_tokens_completion_images
(null)
native_tokens_reasoning
0
native_tokens_cached
0
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
"stop"
usage
0.01805
router
(null)
provider_responses
0
id
"516d5fee-583f-4c85-acbd-a89a3a94dc0c"
status
200
is_byok
false
latency
1381
endpoint_id
"a8708fb9-08ea-43bb-a3b1-165c87ef09df"
provider_name
"Amazon Bedrock"
model_permaslug
"anthropic/claude-4.6-opus-20260205"
api_type
"completions"
id
"gen-1770565206-0JZhCg2XN2Fu0SPT06vo"
upstream_id
"516d5fee-583f-4c85-acbd-a89a3a94dc0c"
total_cost
0.01805
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Amazon Bedrock"
Evaluation details
Result
Evaluator
Details
Meta Data
99.3769%
Matches word count
n/a
neededClean
false
words
505
43.8700%
Dialogue to Total Word Ratio
Ratio: 34.06%, Deviation: 4.06%
neededClean
false
wordsTotal
505
wordsDialogue
172
71.6235%