NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYCPWDF1X8CP03E8BJ9K74H3
Run Details
Model:
Claude Opus 5
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
22.4s
Cost:
$0.031265
Provider:
Amazon Bedrock
Evaluation summary
Score
Evaluator
Details Link
6.09%
Matches word count
(details)
4.94%
Dialogue to Total Word Ratio
(details)
5.5135%
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
The van smelled of stale crisps and ambition. Molly had spread a napkin across the dashboard and was drawing on it with a stub of eyeliner pencil. "Right," she said. "This is the bouncy castle." "That's a square." "It's a castle." "Castles have turrets, Molly. Battlements. That's a square with a smaller square inside it." "The smaller square is Kevin." Bill leaned in, squinting. "Who's Kevin?" "The dad. Party dad. Owns the house, owns the cake, owns nine hundred pounds of unopened presents currently stacked on a trestle table in the conservatory." "Nine hundred?" "It's Persephone's seventh birthday. Her father does something unforgivable with hedge funds. The other parents are competing." Molly tapped the napkin. "There's a child in that house about to receive her fourth scooter." "Fourth." "She won't miss three of them." Bill sat back and rubbed his face with both hands, which was what he did instead of praying. "We're robbing a seven-year-old." "We're robbing a hedge fund manager's insurance policy. Persephone gets a cheque and a lesson about impermanence." "You're a monster." "I'm a redistributor." She held up the napkin. "Now. Conservatory. Sliding door, brass handle, lock so old it's practically nostalgic. Forty seconds." "And the guests?" "Twenty-two children, eleven adults, and an entertainer called Mr Wobbles." "Mr Wobbles." "Balloon animals, some light ventriloquism, twelve-thirty to one-fifteen. Which is when every adult in that house will be standing in the lounge holding a warm glass of prosecco pretending to enjoy a puppet." "And where will we be?" "In the conservatory, loading a wheelbarrow." "We have a wheelbarrow?" "Kevin has a wheelbarrow. It's beside the composter. I've seen it on the Zoopla photos." Bill was quiet a moment. "You looked up the house on Zoopla." "Sold in 2019 for one-point-two. Kitchen's been extended since." She folded the napkin and tucked it in her bra, which she believed was a safe. "Your job is the gate. There's a latch, it sticks, and you're going to hold it open with the good-natured air of a man whose nephew is inside." "I don't have a nephew." "Today you have a nephew called Toby and he's allergic to almonds. If anyone talks to you, lead with the almonds. It ends conversations." "And if it doesn't?" "Then talk about parking." Bill nodded slowly. "One problem." "Go on." "You've built a plan that hinges entirely on a man named Mr Wobbles holding the attention of twenty-two children for forty-five minutes. Molly, I've seen children. They're feral. They'll last nine minutes and then they'll be everywhere, like water." Molly stared through the windscreen at the house, at the bunting sagging between two hanging baskets. "That," she said, "is why we're not stealing at twelve-thirty." "We're not?" "We're stealing at one-forty. Cake." She turned to him, radiant. "Bill, nobody has ever left a room while a cake was being carried into it. Candles, lights off, singing. Everyone films it. Everyone. Eleven adults, eleven phones, all pointed at a single flaming sponge." Bill considered this for a long time. "And Mr Wobbles?" "Mr Wobbles," said Molly, checking her watch, "is being paid two hundred pounds to sing very loudly."
created_at
"2026-07-25T13:19:24.648Z"
7/25/2026, 1:19:24 PM
model
"anthropic/claude-opus-5-20260723"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2127
moderation_latency
(null)
generation_time
22283
tokens_prompt
75
tokens_completion
832
native_tokens_prompt
93
native_tokens_completion
1232
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
0.031265
router
(null)
provider_responses
0
endpoint_id
"76cb4608-f48c-483d-8da8-9957fb44244e"
id
"msg_011CdNqqZhwvEH85drGtoYm7"
is_byok
false
latency
2127
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-1784985564-RpuhIVo31NLrJS8dnqDA"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1784985564-BsbiUdQwgW2NxMc0a5Fn"
upstream_id
"msg_011CdNqqZhwvEH85drGtoYm7"
total_cost
0.031265
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
6.09%
Matches word count
n/a
neededClean
false
words
523
4.94%
Dialogue to Total Word Ratio
Ratio: 77.76%, Deviation: 7.76%
neededClean
false
wordsTotal
535
wordsDialogue
416
5.5135%