NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYCQS2PTW3FYWCW8BA7X5R4T
Run Details
Model:
Claude Opus 5 (Reasoning, Low)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
24.4s
Cost:
$0.032815
Provider:
Amazon Bedrock
Evaluation summary
Score
Evaluator
Details Link
43.38%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
21.6892%
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 place mat in three colors of crayon, because the diner had run out of pens and Molly had run out of patience. "That's the bouncy castle," she said, tapping a purple blob. "That's the piƱata. And that, Bill, is our exit." Bill leaned in, squinting at a green scribble that might have been a hedge or a very large dog. He had been a professional for nineteen years. He had cracked a safe in Antwerp with nothing but a stethoscope and a head cold. He was now being briefed on a job involving a bouncy castle. "Explain to me again," he said, "why we are robbing a seven-year-old." "We're not robbing a seven-year-old. We're robbing the *guests* of a seven-year-old, which is entirely different, morally speaking, and also the guests are the children of the four wealthiest families in the county." Molly stole a triangle of his toast. "Tabitha Pemberton-Voss. Her father imports marble. Last year, a child gave her a pony. Wrapped." "How do you wrap a pony?" "Badly. That's not the point." The point, she explained through a mouthful of toast, was the present table. Every gift arrived unopened and was stacked in the conservatory, unguarded, while thirty-one children were funneled into the garden for ninety minutes of organized activities by a woman named Fiona who ran her entertainments like a Prussian field marshal. Ninety minutes. Two adults. One wheelbarrow, disguised, Molly said, as a wheelbarrow. Bill set down his coffee. "And the parents?" "Drinking. Loudly. In the marquee. They've hired a man to make cocktails out of gin and something called elderflower cordial, and by two o'clock they will not be able to identify their own children, let alone us." "Us being." "You're the magician." Molly slid a photograph across the table: a gaunt man in a plum waistcoat, mid-flourish, doves erupting from a hat with visible reluctance. "Marvo the Mysterious. Cancelled this morning. Food poisoning. Prawns." "You poisoned a children's entertainer." "I bought him prawns. What he did with them was his own business." She produced the plum waistcoat from a carrier bag, and Bill regarded it the way a man regards a diagnosis. "You go in the front. You do fifteen minutes of tricks, badly, so nobody asks for more. Then you tell them the finale requires absolute silence and a closed conservatory door." "And you?" "I'm the pony girl. Nobody questions a woman leading a pony. I come round the side, load the barrow, walk out through the tradesmen's gate at half past three whistling something wholesome." Bill picked up the waistcoat. It smelled faintly of prawns already, which he chose not to investigate. Nineteen years, he thought. Antwerp. The stethoscope. The head cold. "What's my magician's name?" "Bilbonzo." "Absolutely not." "Bilbonzo the Astounding." "Molly." "It's on the poster already," she said, standing, and dropped a five-pound note on the table for a bill of eleven pounds forty. "Wear the hat. Children love a hat." Bill wore the hat. It fit, which he found the most upsetting part of the entire enterprise.
created_at
"2026-07-25T13:35:03.905Z"
7/25/2026, 1:35:03 PM
model
"anthropic/claude-opus-5-20260723"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3873
moderation_latency
(null)
generation_time
24312
tokens_prompt
75
tokens_completion
889
native_tokens_prompt
93
native_tokens_completion
1294
native_tokens_completion_images
(null)
native_tokens_reasoning
103
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
0.032815
router
(null)
provider_responses
0
endpoint_id
"76cb4608-f48c-483d-8da8-9957fb44244e"
id
"msg_011CdNs2oNtga14x6fRgRLoe"
is_byok
false
latency
1270
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-1784986503-YMsXt00tGnPXiP7bKk7S"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1784986503-TONGMTBQwT7Px6OKQB5b"
upstream_id
"msg_011CdNs2oNtga14x6fRgRLoe"
total_cost
0.032815
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
43.38%
Matches word count
n/a
neededClean
false
words
517
0.00%
Dialogue to Total Word Ratio
Ratio: 47.25%, Deviation: 17.25%
neededClean
false
wordsTotal
527
wordsDialogue
249
21.6892%