NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYCR3M0SBAK3ZDQBCNS9T81G
Run Details
Model:
Claude Opus 5 (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
21.9s
Cost:
$0.029665
Provider:
Amazon Bedrock
Evaluation summary
Score
Evaluator
Details Link
97.63%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
48.8138%
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 blueprint across the hood of the van, smoothing it with the side of her hand. "That's a napkin," Bill said. "It's a *tactical* napkin. Now pay attention. This is the Hendersons' back garden. Bouncy castle here, buffet table here, and the presents — the presents, Bill — are stacked on a trestle table beneath the gazebo." "How many?" "Thirty-one children were invited. Assume one gift each, plus the grandparents overcompensating. Call it forty units." "Forty units of what, though? That's what worries me. Could be forty jigsaw puzzles. Could be forty pairs of novelty socks." "It's a seventh birthday, Bill. Seven is prime real estate. Seven is scooters and tablets and those little robot dogs that cost more than my rent." Bill scratched his jaw. "And the security?" "One clown." "One clown." "Name of Bingo. Forty-two, divorced, does birthdays and the occasional bar mitzvah. Weak left knee." "How do you know about his knee?" "I have a source." "Molly." "I dated him for six weeks in 2019 and he cried during *Paddington 2*, which is completely reasonable, but he also told me his medical history over starters, which is not." Bill folded his arms. "And the parents? There's always a father with a camcorder." "Gareth. He'll be at the barbecue from noon onward, entirely absorbed in a conversation about decking. Nothing short of a house fire moves a man off decking. The mother, Alison, will be inside slicing a cake shaped like a dinosaur. That leaves the garden to Bingo and thirty-one sugar-drunk children." "That's the part I object to." "Which part?" "The thirty-one children. They're a swarm, Molly. They have no fear and no fixed schedule. A child will look you dead in the eye and ask why your face is like that." "Which is why you're going in as the balloon man." Bill went very still. "Say that again." "Bill Baxter, professional balloon artist. You've got the apron, you've got the pump, you've got a laminated badge I made at the library for eighty pence." "I can't make balloon animals." "You made a swan at Deb's engagement do." "That was an accident and everyone agreed it looked like a hernia." "Then make hernias. Children have no standards, they only have volume. You keep them clustered by the trampoline while I reverse the van up the side alley and load the gazebo table into the back. Ninety seconds, tops." "And if one of them follows you?" "Then you say the magic words." "What magic words?" "'Who wants to see Bingo do a cartwheel?'" Bill exhaled slowly through his nose. "The knee." "The knee," Molly agreed. "Chaos. Ambulance, probably. Nobody's counting presents while a clown is being stretchered past the trifle." "You're a monster." "I'm an opportunist. There's a difference, and it's mostly paperwork." She folded the napkin and tucked it into her sleeve. "Party starts at two. Be at the gate at ten past, and Bill — smile. You're a balloon man now." "I hate balloon men." "Everybody hates balloon men. That's what makes it perfect."
created_at
"2026-07-25T13:40:49.313Z"
7/25/2026, 1:40:49 PM
model
"anthropic/claude-opus-5-20260723"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2993
moderation_latency
(null)
generation_time
21809
tokens_prompt
75
tokens_completion
830
native_tokens_prompt
93
native_tokens_completion
1168
native_tokens_completion_images
(null)
native_tokens_reasoning
53
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
0.029665
router
(null)
provider_responses
0
endpoint_id
"76cb4608-f48c-483d-8da8-9957fb44244e"
id
"msg_011CdNsUGE844pEx5g9ASB2C"
is_byok
false
latency
1276
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-1784986849-ib781oTRfKUIsjO1VJcF"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1784986849-Q1XNZnynpAE2StWYILJR"
upstream_id
"msg_011CdNsUGE844pEx5g9ASB2C"
total_cost
0.029665
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
97.63%
Matches word count
n/a
neededClean
false
words
507
0.00%
Dialogue to Total Word Ratio
Ratio: 90.23%, Deviation: 20.23%
neededClean
false
wordsTotal
512
wordsDialogue
462
48.8138%