NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYCRKPG3G3J37M81YENJY2N8
Run Details
Model:
Claude Opus 5 (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
22.9s
Cost:
$0.028815
Provider:
Amazon Bedrock
Evaluation summary
Score
Evaluator
Details Link
20.19%
Matches word count
(details)
90.99%
Dialogue to Total Word Ratio
(details)
55.5908%
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 blueprints were drawn on the back of a pizza box, which Bill considered the height of professionalism, since the last job had been planned on a napkin that dissolved in the rain. Molly leaned over the table, tapping a greasy corner with her pen. The Hendersons' backyard had been rendered in ballpoint with surprising care: the inflatable castle, the trestle table sagging under a sheet cake, and there, in the northeast corner, shaded with obsessive crosshatching, the Pile. "Forty-two children," she said. "Forty-two presents. Possibly more, because Cassandra Henderson invited the entire second grade and those people are competitive gifters." Bill scratched his beard with the blunt end of a candy cane he'd been saving since December. He had the sunken look of a man who had spent his morning learning to twist balloon animals off a library DVD and had produced only what he described as sausages with regret. "I want it on the record," he said, "that I raised concerns." "You raised eleven concerns." "And I'm raising a twelfth. It's a child's birthday, Molly. There's a moral component." Molly straightened and gave him the look she reserved for men who developed ethics eight hours before a job. Bill withered gracefully, as he did most things. The plan, as she explained it, was elegant in the way that a mousetrap is elegant. Bill would arrive at eleven-forty in the rented clown suit, which smelled faintly of another man's despair, and introduce himself as Bobo, sent by the agency. There was no agency. There was, however, a laminated business card Molly had made at a print shop in a strip mall, and laminate, she had learned over nineteen years in the trade, was the closest thing the world had to a passport. At noon, Bobo would begin the balloon portion of the entertainment, herding all forty-two children toward the far hedge under the pretense of a game called Silent Hippopotamus. "They won't be silent," Bill said. "They'll be six. Tell them the winner gets cake first and they'll behave like Trappist monks." Meanwhile Molly, in a floral sundress and enormous sunglasses, would drift among the mothers as Aunt Deborah from Sacramento, a woman so vaguely related that no one would risk asking. She would carry a canvas tote large enough to smuggle a labrador. Two minutes at the Pile, working from the outside in, largest boxes first. Bill's cousin Denny would idle the minivan at the alley gate with the sliding door open and, God willing, the radio off. "And if someone asks what I'm doing?" "You say you're moving the presents inside because of the wasps." Bill considered this. "There are wasps?" "Bill." Molly folded the pizza box in half with the tenderness of a general burning maps. "There are always wasps. Nobody in the history of the American backyard has ever contested a wasp." He nodded slowly, then reached for the clown wig, held it up to the window light, and sighed at it the way a knight might sigh at a rusted helmet. "For the record," he said again. "Noted," said Molly. "Get in the car."
created_at
"2026-07-25T13:49:36.137Z"
7/25/2026, 1:49:36 PM
model
"anthropic/claude-opus-5-20260723"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3180
moderation_latency
(null)
generation_time
22838
tokens_prompt
75
tokens_completion
826
native_tokens_prompt
93
native_tokens_completion
1134
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
0.028815
router
(null)
provider_responses
0
endpoint_id
"76cb4608-f48c-483d-8da8-9957fb44244e"
id
"msg_011CdNt96KenNBLVJitiP33Q"
is_byok
false
latency
1120
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-1784987376-lLJninEop3xzeCCjVrY6"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1784987376-ZXmhitK3jJPcFJJ9mUmZ"
upstream_id
"msg_011CdNt96KenNBLVJitiP33Q"
total_cost
0.028815
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
20.19%
Matches word count
n/a
neededClean
false
words
520
90.99%
Dialogue to Total Word Ratio
Ratio: 28.63%, Deviation: 1.37%
neededClean
false
wordsTotal
524
wordsDialogue
150
55.5908%