NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYCPRD0RT0JHE43DHP4EVFPB
Run Details
Model:
Claude Opus 5
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
21.0s
Cost:
$0.027815
Provider:
Amazon Bedrock
Evaluation summary
Score
Evaluator
Details Link
35.00%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
17.5012%
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 rain had turned the alley behind Kowalski's Bakery into a soup of flour dust and cigarette ends, which suited Molly fine because nobody sensible loitered there, and she and Bill had never once been accused of sense. She had spread the blueprints across a milk crate, weighing down the corners with a hubcap, a half-eaten cannoli, and Bill's left shoe. "That's the bouncy castle," she said, tapping the paper. "Inflatable. Load-bearing, if you believe the manual, which I don't." Bill leaned in, squinting through the drizzle. His sock had gone dark at the toe. "And the presents?" "Gift table. North wall. Under the piñata." He straightened up with the wounded air of a man who had discovered a moral line and found himself on the correct side of it for once. "Molly. It's a birthday party." "It's a *seventh* birthday party, Bill. Seventh. The gift ceiling on a seventh is astronomical. Bikes. Scooters. Those little tablets they give children now so they'll stop asking questions." She rolled the blueprint into a tube and pointed it at him like a sword. "Do you know what a bicycle fetches at the flea market in April? April, Bill. Bicycle season." "They're *children's* bicycles." "Children are the fastest-growing demographic in the bicycle sector." Bill wiped rain off his forehead and left a streak of alley grime in its place. Twenty-two years they'd been at this, and he still had the face of a man who wanted to be talked out of things. That was why she kept him. A conscience was useful in a partner, provided you could out-talk it, and Molly could out-talk a lawyer with a head cold. "There'll be a clown," he said, hopefully. "There will be a clown." "I don't like clowns." "Nobody likes clowns, Bill, that's the whole architecture of the clown. They're built out of nobody liking them." She unrolled the blueprint again and jabbed at a rectangle labeled *utility*. "Two thirty-five, the magician does his set. Everyone faces east. You come through the side gate dressed as the man from the balloon company." "I don't have a balloon company outfit." "You have a yellow polo shirt." "That's my funeral polo." "It'll be a funeral for the piñata." She grinned at him, and he did not grin back, and she carried on undeterred, which was, broadly, the story of their friendship. "You go for the table. I'll be inside the castle." Bill blinked. "You'll be *inside* the bouncy castle." "Interior. Uncontested airspace. Nobody looks in a bouncy castle for an adult woman." "Because there's never been one in there." "Precisely. Unprecedented. Unsearchable." She folded the plans into her coat and retrieved his shoe from the crate, handing it over with the solemnity of a knighthood. "Two thirty-five, Bill. Yellow polo. And whatever happens, do not, under any circumstances, engage with the clown." He put the shoe on, standing storklike in the puddle, and sighed the sigh of a man already imagining the courtroom. "What happens if I engage with the clown?" Molly was already walking away. "Then we're both going to prison in balloon hats."
created_at
"2026-07-25T13:17:13.12Z"
7/25/2026, 1:17:13 PM
model
"anthropic/claude-opus-5-20260723"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1552
moderation_latency
(null)
generation_time
20821
tokens_prompt
75
tokens_completion
780
native_tokens_prompt
93
native_tokens_completion
1094
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
0.027815
router
(null)
provider_responses
0
endpoint_id
"76cb4608-f48c-483d-8da8-9957fb44244e"
id
"msg_011CdNqfsLZrNa3W6BSxkYso"
is_byok
false
latency
1552
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-1784985433-RQJXLrAVnr1mtCflid3B"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1784985433-zlCFFcI35nj0WV5ifsBs"
upstream_id
"msg_011CdNqfsLZrNa3W6BSxkYso"
total_cost
0.027815
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
35.00%
Matches word count
n/a
neededClean
false
words
518
0.00%
Dialogue to Total Word Ratio
Ratio: 46.37%, Deviation: 16.37%
neededClean
false
wordsTotal
524
wordsDialogue
243
17.5012%