NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYCPP5EK0NDBX1MZY58XDF3K
Run Details
Model:
Claude Opus 5
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
22.2s
Cost:
$0.027190
Provider:
Amazon Bedrock
Evaluation summary
Score
Evaluator
Details Link
86.38%
Matches word count
(details)
63.58%
Dialogue to Total Word Ratio
(details)
74.9787%
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 nacho cheese and ambition. Bill had parked it beneath a sagging elm across the street from 44 Winslow Lane, where pink and yellow balloons strangled a mailbox and a bounce house shuddered in the breeze like a lung. Molly lowered the binoculars. She had stolen them from a birdwatcher in Devon, which she considered a victimless crime because the man had been rude about her hat. "Twenty-three children," she said. "One clown. Two grandmothers, minimum. That's the problem." "Grandmothers?" "Grandmothers see everything, Bill. They have no hobbies but observation." Bill was eating a sandwich he'd been eating since Tuesday. He chewed thoughtfully and squinted at the house, where a folding table groaned under a pyramid of wrapped boxes, each one glittering like a small crime waiting to happen. Somewhere in that stack, the internet had assured them, was a limited-edition Battle Falcon playset worth eleven hundred pounds to a collector in Rotterdam. The rest was gravy. Molly loved gravy. "So we go in through the side gate," Bill said. "During cake." "During cake nothing. Cake is when they're all looking in one direction, sure, but they're looking at the *table*. Cake happens at the table. The presents are at the table." "Then when?" "Piñata." Bill stopped chewing. It was, he had to admit, magnificent. Piñata meant blindfolds. Piñata meant chaos and swinging and every adult in the garden shouting contradictory directions at a small girl with a stick. Piñata meant nobody, not one soul, watching the pile. "You're a genius," he said. "I'm a professional. There's a difference, but I accept the compliment." They worked out the details on the back of a parking ticket. Bill would go over the fence at the rear, cross the patio, and load the presents into two large black rubbish sacks, which they would carry out through the side gate looking purposeful. Nobody stops a man carrying rubbish. It is the closest thing to invisibility available at street level. "And if someone asks?" "You're the man from the council." "On a Saturday?" "Bill, you own a high-visibility vest. That vest is a passport to every locked door in this country." At 2:40 the clown emerged and honked something that might have been a horn or a lung. At 2:55 the singing started. Molly checked her watch and felt the old warmth in her chest, the one that came just before things went beautifully wrong. Then a woman in a cardigan carried the entire pile of presents indoors, one armful at a time, humming. Bill lowered the sandwich. "She's moving them." "I can see that." "Inside. She's putting them inside the house." "I can *see* that, Bill." The piñata rose over the garden like a small papier-mâché moon. The children shrieked. The clown honked. And Molly, watching the last present vanish through the patio door, tapped her fingers on the dashboard and began, slowly and with a certain grim delight, to rethink absolutely everything. "Alright," she said. "New plan. How do you feel about clowns?" Bill had already begun to sweat.
created_at
"2026-07-25T13:15:59.832Z"
7/25/2026, 1:15:59 PM
model
"anthropic/claude-opus-5-20260723"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2031
moderation_latency
(null)
generation_time
22061
tokens_prompt
75
tokens_completion
782
native_tokens_prompt
93
native_tokens_completion
1069
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
0.02719
router
(null)
provider_responses
0
endpoint_id
"76cb4608-f48c-483d-8da8-9957fb44244e"
id
"msg_011CdNqaTpoxDVTY8aiZQFsL"
is_byok
false
latency
2031
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-1784985359-jGWro55BramMhv5ZLdNK"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1784985359-3b763utLGMmQdCzWEPX2"
upstream_id
"msg_011CdNqaTpoxDVTY8aiZQFsL"
total_cost
0.02719
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
86.38%
Matches word count
n/a
neededClean
false
words
511
63.58%
Dialogue to Total Word Ratio
Ratio: 26.99%, Deviation: 3.01%
neededClean
false
wordsTotal
515
wordsDialogue
139
74.9787%