NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYCPNFFNWQZPMBFS6F0EG24F
Run Details
Model:
Claude Opus 5
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
20.2s
Cost:
$0.025740
Provider:
Amazon Bedrock
Evaluation summary
Score
Evaluator
Details Link
99.74%
Matches word count
(details)
0.07%
Dialogue to Total Word Ratio
(details)
49.9055%
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 popcorn and ambition. Molly had parked it beneath a drooping willow across from 14 Marigold Lane, where a rented bouncy castle wobbled like a fat man laughing. Bill was in the passenger seat, eating a sandwich he had not brought and could not account for, studying the house through a pair of opera glasses he'd stolen from a production of *La Traviata* in 1998. "Nine presents on the trestle table," he said. "Ten if you count the envelope, which I don't, because envelopes are cowardice." "Count the envelope." "Envelopes are what you give when you've forgotten. There's no soul in an envelope." Molly unfolded the floor plan she'd sketched on the back of a takeaway menu. She had labeled the kitchen KITCHEN, the lounge LOUNGE, and the downstairs toilet CHOKE POINT. The birthday child was one Arthur Pemberton-Sneed, aged seven, whose parents had booked a magician, a pony, and a caterer specializing in what the invitation described as "artisanal party rings." "Entry through the side gate at fourteen hundred hours," she said. "Magician goes on at two. Every adult in that garden will be watching a man pull a scarf out of another man's ear. That's our window." "Twelve minutes." "Nine. Magicians overrun. It's the one thing you can rely on in this world." Bill lowered the opera glasses. There was mustard on his thumb, which he addressed with the fastidiousness of a surgeon. Twenty-two years they had worked together, since the great Bakewell tart incident, and in all that time Molly had never known him to attempt a job sober, hungry, or without a wholly unnecessary hat. Today's was a fedora. It made him look like a man who had once been told he looked good in a fedora and had believed it. "There's a dog," he said. "There's always a dog." "A big one. Some manner of retriever. It's already eaten a paper plate." "Then it'll eat the sausage rolls I've got in the glovebox and it'll love us like family." She tapped the menu. "You take the table. I take the pile by the French doors — that's where the good stuff goes, the aunties always put theirs by the doors because they arrive last and can't be bothered to walk. Bags open, silent, no shaking. If you shake a present you'll want to know what's inside and then you'll get sentimental." "I don't get sentimental." "You wept over a Meccano set in Rochdale." "That was a *building* set, Molly. That was infrastructure." She let it pass. Across the road, a woman in a sunhat was carrying out a cake shaped, improbably, like a submarine. Behind her came a pony, led by a teenager who had clearly been told this would be a summer job with prospects. "One rule," Molly said, starting the engine and immediately turning it off again, which was a ritual, not a mistake. "We do not touch the cake." Bill adjusted his fedora and said nothing, which was, in its way, an answer.
created_at
"2026-07-25T13:15:37.338Z"
7/25/2026, 1:15:37 PM
model
"anthropic/claude-opus-5-20260723"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1584
moderation_latency
(null)
generation_time
20083
tokens_prompt
75
tokens_completion
729
native_tokens_prompt
93
native_tokens_completion
1011
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
0.02574
router
(null)
provider_responses
0
endpoint_id
"76cb4608-f48c-483d-8da8-9957fb44244e"
id
"msg_011CdNqYoMktoQnxJiczGLD9"
is_byok
false
latency
1584
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-1784985337-2XHSJoIHcEKXRZ1frGjV"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1784985337-havgTmrPEz0eEIH7P2Bj"
upstream_id
"msg_011CdNqYoMktoQnxJiczGLD9"
total_cost
0.02574
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
99.74%
Matches word count
n/a
neededClean
false
words
504
0.07%
Dialogue to Total Word Ratio
Ratio: 42.09%, Deviation: 12.09%
neededClean
false
wordsTotal
506
wordsDialogue
213
49.9055%