NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYCPWJMPZQWR6JZNVAPZ20GG
Run Details
Model:
Claude Opus 5
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
23.0s
Cost:
$0.028740
Provider:
Amazon Bedrock
Evaluation summary
Score
Evaluator
Details Link
86.38%
Matches word count
(details)
0.02%
Dialogue to Total Word Ratio
(details)
43.2022%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 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
The van smelled of cold pizza and ambition. Molly had spread a hand-drawn map across the dashboard, weighted at the corners with a stapler, a lighter, and two petrified doughnuts. "Right," she said. "Bouncy castle here. Trestle table with the sandwiches here. And the presents—" she stabbed the map, "—here. Under the window. Stacked like a pirate hoard." Bill leaned in, squinting. "Why's the bouncy castle drawn bigger than the house?" "Because it's important, Bill. It's tactically important." "It's a castle for toddlers." "It's cover. It's soft cover. Nobody looks at a man crawling behind a bouncy castle." "Everybody looks at a man crawling behind a bouncy castle." Bill sat back and folded his arms over his stomach, which had recently acquired opinions of its own. "That's the sort of thing people film." Molly ignored him and drew an arrow. "Fourteen children. One mother, one grandmother, one dad who'll be at the barbecue, unreachable, spiritually. And a magician." "A magician." Bill's voice went flat. "You didn't say magician." "Mister Bamboozle." "Molly." "He's seventy-one and he does the same six tricks." "Magicians notice things," Bill said. "That's the whole job. That's literally the entire job description. Noticing things and lying about it." "He's a retired plumber from Dagenham who owns a dove." "A dove is a witness." Molly rolled up the map with the crispness of a general who had lost patience with her cavalry. "Half two, the cake comes out. Candles. Lights off. Everyone singing. That's ninety seconds where fourteen children and three adults are staring at a burning sponge. That's the window." "Ninety seconds to move forty presents." "Thirty-eight. I counted from the fence." "You counted presents through a fence?" "I've got binoculars. Don't make it weird." Bill rubbed his face. "And the bin bags?" "Two each. Black, not the green ones, the green ones look like garden waste and somebody helpful will put them out." "That happened once." "It happened twice, Bill, and the second time was a wedding." He let that lie between them for a moment, because it was true and it still stung. Then he said, "What about the dog?" "There's no dog." "There's a bowl. On the step. I saw a bowl on the step." Molly went very still, the way she did when the plan developed a seam. "It could be a cat bowl." "It said *Sausage*." "Cats can be called Sausage." "Cats are called Mittens and Duchess and Lord Byron. Dogs are called Sausage." Bill turned to look at her properly. "Molly, tell me you didn't drive us across two counties to rob a house with an unaccounted-for Sausage in it." "I'll bring ham." "You'll bring ham." "Everybody's got a price, Bill. His is ham." Bill exhaled, long and slow, and reached for a petrified doughnut, because a man had to eat something before ruining a child's afternoon. Through the windscreen, across the road, a red balloon escaped the gatepost and rose, unhurried, into a blameless blue sky. "Half two," he said. "Half two," said Molly. "Wear the party hat. It's respectful."
created_at
"2026-07-25T13:19:29.948Z"
7/25/2026, 1:19:29 PM
model
"anthropic/claude-opus-5-20260723"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2364
moderation_latency
(null)
generation_time
22878
tokens_prompt
75
tokens_completion
788
native_tokens_prompt
93
native_tokens_completion
1131
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)
num_fetches
(null)
web_search_engine
(null)
origin
"https://ncbench.com/"
is_byok
false
finish_reason
"stop"
native_finish_reason
"end_turn"
service_tier
"default"
usage
0.02874
router
(null)
provider_responses
0
endpoint_id
"76cb4608-f48c-483d-8da8-9957fb44244e"
id
"msg_011CdNqqyMFTWUJKikxDtbtx"
is_byok
false
latency
2364
model_permaslug
"anthropic/claude-opus-5-20260723"
provider_name
"Amazon Bedrock"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.18.0; linux; x64))"
http_referer
(null)
request_id
"req-1784985569-pP3nZKlIE7NvRney6ta0"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1784985569-3terN2PczmwbMicPN0BZ"
upstream_id
"msg_011CdNqqyMFTWUJKikxDtbtx"
total_cost
0.02874
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Amazon Bedrock"
response_cache_source_id
(null)
data_region
"global"
Evaluation details
Result
Evaluator
Details
Meta Data
86.38%
Matches word count
n/a
neededClean
false
words
511
0.02%
Dialogue to Total Word Ratio
Ratio: 62.91%, Deviation: 12.91%
neededClean
false
wordsTotal
515
wordsDialogue
324
43.2022%