NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYCQR4C8XQTCGE00CWR6H5T4
Run Details
Model:
Claude Opus 5 (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
22.8s
Cost:
$0.031115
Provider:
Amazon Bedrock
Evaluation summary
Score
Evaluator
Details Link
90.48%
Matches word count
(details)
98.53%
Dialogue to Total Word Ratio
(details)
94.5055%
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 liquorice and ambition. Molly had spread the blueprints across the dashboard — technically not blueprints, technically a leaflet for Wobbles the Clown's Deluxe Fun Package, but she'd drawn arrows on it in red biro and that made it strategy. Bill was eating a sausage roll with the reverence of a man attending his own christening. Flakes of pastry settled into the folds of his stolen fleece, which read STAFF in cheerful yellow letters, and which he had not once removed since acquiring it in 2019. "Present table's here," Molly said, tapping a jam-coloured smudge. "Beside the bouncy castle. Forty-two guests, average gift value eleven quid, plus whatever the grandparents cough up." "Grandparents give money." "Grandparents give *cash*, Bill. Which is the finest gift of all, and one we shall be honouring." He nodded, chewing. Somewhere behind his eyes, a small abacus was clacking. Beyond the windscreen, number fourteen Marlowe Crescent was hemorrhaging balloons. A woman in a linen jumpsuit was zip-tying bunting to a lamppost with the grim efficiency of someone who had been awake since four. Children were arriving in shoals, each clutching a rectangle of wrapping paper, each rectangle representing — in Molly's calculations — approximately four pounds sixty after fencing. "There's a moral dimension here," Bill said. "There is not." "There's a *bit* of one." Molly turned to face him fully, which in the confines of a Vauxhall Combo was a considerable manoeuvre. She let the silence do most of the work, the way her mother had. "Last April," she said, "you took a charity tin off a bar in Streatham. It was shaped like a dog." "It was shaped like a *cat*, and I've made peace with that." "So we're peaceful. Good." She smoothed the leaflet. "Fourteen minutes. That's the window. Wobbles does his balloon-animal segment at half two, and every adult in that garden will be filming a poodle being twisted into existence. Nobody films a present table. Nobody has *ever* filmed a present table." Bill considered this and found it philosophically sound, even beautiful. He wiped his hands on his knees. The plan, such as it was, involved two wheelie bins, a laundry hamper, and Bill's unshakeable belief that a man in a STAFF fleece could walk into any building in Britain and be offered a cup of tea. Molly's role was to hover near the trestle table with a bin bag and the expression of a woman conscientiously tidying up, an expression she had practised in the bathroom mirror for eleven minutes that morning and now believed to be her best work. "And if someone asks?" "We're from the hire company. Collecting the gift receptacles." "That's not a thing." "It's a thing if you say it flatly enough." He looked out at the balloons, at the bunting, at the terrible innocence of it all, and felt something shift pleasantly into place, the way a key does. "Fourteen minutes," he said. Molly folded the leaflet into her sleeve, and they got out of the van as though the day belonged to them.
created_at
"2026-07-25T13:34:32.848Z"
7/25/2026, 1:34:32 PM
model
"anthropic/claude-opus-5-20260723"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
4531
moderation_latency
(null)
generation_time
22742
tokens_prompt
75
tokens_completion
855
native_tokens_prompt
93
native_tokens_completion
1226
native_tokens_completion_images
(null)
native_tokens_reasoning
81
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
0.031115
router
(null)
provider_responses
0
endpoint_id
"76cb4608-f48c-483d-8da8-9957fb44244e"
id
"msg_011CdNrzXUNbZd7XQV6o6DbZ"
is_byok
false
latency
1478
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-1784986472-NOuVcSOOqpRHQ5zXUOa5"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1784986472-vKZVQpzZdzqBJdVNWJZF"
upstream_id
"msg_011CdNrzXUNbZd7XQV6o6DbZ"
total_cost
0.031115
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
90.48%
Matches word count
n/a
neededClean
false
words
510
98.53%
Dialogue to Total Word Ratio
Ratio: 30.54%, Deviation: 0.54%
neededClean
false
wordsTotal
514
wordsDialogue
157
94.5055%