NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYCR4BHA2K2GKRSYTYJ3S98B
Run Details
Model:
Claude Opus 5 (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
23.9s
Cost:
$0.031440
Provider:
Amazon Bedrock
Evaluation summary
Score
Evaluator
Details Link
2.01%
Matches word count
(details)
0.04%
Dialogue to Total Word Ratio
(details)
1.0282%
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 spread the napkin across the diner table and smoothed it flat like a blueprint, which, in fairness, it was. "Okay. Ravensworth Estates, sixteen Fielding Court. Birthday girl's name is Poppy. She's turning seven." "Seven," Bill said. "Is that a problem?" "It's just that seven is a big one. Seven's when the presents get good. Bicycles. Scooters. That trampoline nonsense." "Which is precisely why we're going Saturday and not, say, hitting a christening." Bill dunked a chip in gravy and held it aloft, dripping. "I want it noted that I raised concerns." "Noted. Now, the gift table is here—" she stabbed the napkin, tearing it slightly, "—under the marquee, west side of the garden. Unattended for approximately forty minutes while the entertainer does his set." "There's an entertainer?" "Mr. Wobbles. Balloon animals, close-up magic, allegedly a doves finale." "Doves." Bill sat back. "Molly, we can't compete with doves. Kids love doves." "We're not competing with him, we're *using* him. Everyone's eyes go to the doves, our eyes go to the loot. That's forty minutes of total ocular monopoly." "And how do we get in? It's a garden party. There'll be mums. Molly, you don't know mums like I know mums. A mum can clock a stranger at ninety feet through a hedge while breastfeeding." "That's why you're going in as Uncle Gary." Bill put down the chip. "Whose Uncle Gary?" "Nobody's. There's always an Uncle Gary. Big fella, slightly damp shirt, stands by the crisps, no one's entirely sure whose side he's from. He is the most invisible man in England." "I don't love that you thought of me for that." "I thought of you for that because you have a natural gift, Bill, and you should be flattered." "I'm going to need a lanyard or something. A prop." "You'll have a card. A big novelty card, forty quid, from The Range. You hold it under your arm the entire time and nobody questions you, because a man holding a big card is a man who has been invited." Bill considered this and, visibly, could not fault it. "And you?" "Catering." "You can't cook." "I don't need to cook, I need to carry trays. In and out, kitchen to marquee, twelve times. On the eleventh trip the tray goes out full of presents under a tea towel." "A tea towel." "A tea towel is the greatest cloaking device in the history of British crime. Nobody has ever lifted a tea towel. It's practically a locked door." "And the trampoline? You're carrying a trampoline out under a tea towel, are you?" Molly paused. "The trampoline we leave." "Ah." "We leave the trampoline, Bill, because we are professionals with realistic expectations." "So we're stealing from a seven-year-old, but *considerately*." "We're stealing from her extended family, who spent an average of twenty-two pounds each on rubbish she'll have forgotten by Tuesday. It's practically a favour. Builds character." Bill wiped his hands on his jeans. "One condition." "Go on." "If the doves get loose, I'm keeping one." "Absolutely not." "Then I'm not being Uncle Gary." Molly closed her eyes. "One dove." "Lovely." He raised his mug. "To Poppy."
created_at
"2026-07-25T13:41:13.394Z"
7/25/2026, 1:41:13 PM
model
"anthropic/claude-opus-5-20260723"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3800
moderation_latency
(null)
generation_time
23760
tokens_prompt
75
tokens_completion
852
native_tokens_prompt
93
native_tokens_completion
1239
native_tokens_completion_images
(null)
native_tokens_reasoning
51
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
0.03144
router
(null)
provider_responses
0
endpoint_id
"76cb4608-f48c-483d-8da8-9957fb44244e"
id
"msg_011CdNsW4D7weqa9iQdFqthL"
is_byok
false
latency
1671
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-1784986873-hhUpaXgaztF26BRI3Aot"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1784986873-yiQFns67IM48A4frBURj"
upstream_id
"msg_011CdNsW4D7weqa9iQdFqthL"
total_cost
0.03144
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
2.01%
Matches word count
n/a
neededClean
false
words
525
0.04%
Dialogue to Total Word Ratio
Ratio: 82.42%, Deviation: 12.42%
neededClean
false
wordsTotal
529
wordsDialogue
436
1.0282%