NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYCPKSS7V5KYXSM8Y102H0QB
Run Details
Model:
Claude Opus 5
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
21.5s
Cost:
$0.026490
Provider:
Amazon Bedrock
Evaluation summary
Score
Evaluator
Details Link
68.10%
Matches word count
(details)
30.35%
Dialogue to Total Word Ratio
(details)
49.2237%
Overall
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
The van smelled of stale cheese puffs and ambition. Molly had parked it beneath a sycamore two houses down from the Pemberton residence, where a bouncy castle shaped like a dragon inflated with a wheeze that sounded, to her ears, like the sigh of a dying god. Bill was drawing on the back of a takeout menu, tongue poking from the corner of his mouth. He held the drawing up for inspection: a rectangle, a triangle, and several arrows pointing at nothing in particular. "That's the house?" "That's the house." "Bill, that's a barn." "It's stylized." He turned the menu ninety degrees, considered it, and turned it back. "Point is, the presents go on the trestle table by the patio doors. Kids don't open them till after cake. Cake's at four. We're in and out by three-fifty with eleven, twelve bags of loot." Molly lowered the binoculars. Through the hedge she could see a woman in linen carrying a tray of orange squash, and a man in a paper crown wrestling a piñata into a tree with the grim resolve of someone who had lost an argument that morning and intended to win this. "Twelve bags," she repeated. "Of what, exactly. Talk me through the inventory." "Toys." "Which toys." "The good ones. Kids these days, it's all electronics. Tablets. Those robot dogs. Sold three of those last Christmas for eighty apiece and one of them barked." "One of them barked." "That's a feature." She let the binoculars drop against her chest and rubbed the bridge of her nose. Six years she'd worked with Bill. Six years of drawings that looked like barns, of plans that hinged on the reliable stupidity of strangers, and — she was forced to admit — six years of unbroken success, which she attributed less to Bill's genius than to the fact that nobody ever suspected a man who dressed like a substitute geography teacher. "The father," she said. "Paper crown. He's ex-military." "How do you know that?" "He's tying a knot I've only seen in prison and the Navy." Bill leaned forward and peered through the windscreen, breathing audibly. "He's got soft hands." "He's got a hammer in his belt loop." "For the piñata." "Nobody hammers a piñata, Bill." They sat with that awhile. Somewhere behind the hedge a child began to scream in the high, glorious register that means either delight or grievous injury, and a chorus of other children took up the sound, and then a rendition of something that was not quite Happy Birthday started and immediately collapsed. Molly reached into the back and produced two high-visibility vests and a clipboard, which was, in her professional experience, the closest thing to an invisibility cloak available at cost. "Council," she said. "Wasp inspection. We say there's a nest in the eaves, we say everyone out front for ten minutes, and while they're all standing on the lawn looking up at nothing, you back the van to the patio doors." Bill turned the menu over and began, very slowly, to draw a wasp. "Molly," he said, with feeling. "That's beautiful."
created_at
"2026-07-25T13:14:42.549Z"
7/25/2026, 1:14:42 PM
model
"anthropic/claude-opus-5-20260723"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1632
moderation_latency
182
generation_time
21011
tokens_prompt
75
tokens_completion
764
native_tokens_prompt
93
native_tokens_completion
1041
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
0.02649
router
(null)
provider_responses
0
endpoint_id
"76cb4608-f48c-483d-8da8-9957fb44244e"
id
"msg_011CdNqUnEvat5afg8xiECop"
is_byok
false
latency
1632
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-1784985282-70HiTAesACeDm7mlQpE6"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1784985282-xmkZq8sDcB3PFiqlyj5v"
upstream_id
"msg_011CdNqUnEvat5afg8xiECop"
total_cost
0.02649
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
68.10%
Matches word count
n/a
neededClean
false
words
514
30.35%
Dialogue to Total Word Ratio
Ratio: 34.88%, Deviation: 4.88%
neededClean
false
wordsTotal
516
wordsDialogue
180
49.2237%