NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M37PKPRCN903FNWTS4D12ZCN
Run Details
Model:
Claude Opus 5.5 (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
22.1s
Cost:
$0.032780
Provider:
Claude Platform on AWS
Evaluation summary
Score
Evaluator
Details Link
97.63%
Matches word count
(details)
27.27%
Dialogue to Total Word Ratio
(details)
62.4472%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 50% 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
Molly spread the floor plan across the bonnet of the van and pinned the corners down with four juice boxes. "Right," she said. "Tyler Pemberton. Turning seven. Forty-two guests, one bouncy castle, and a gift table the size of a small hatchback." Bill leaned in until his nose nearly touched the paper. "Why's the garden drawn in crayon?" "Because I copied it off the invitation, Bill. There's a dinosaur on the back." He flipped it over and nodded gravely at the dinosaur, as though it might be in on the job. The van smelled of stale crisps and ambition. Molly had spent twenty years stealing things that mattered: diamonds, a minor Monet, once an entire racehorse that had later bitten her. But these days the museums had lasers and the banks had lawyers, and the only unguarded treasure left in the county was a mountain of wrapped presents in a suburban garden, protected by one exhausted father holding barbecue tongs. "So what's my part?" Bill asked. "You're the entertainment." "I don't do entertainment." "You do today. You're Mister Giggles, the clown." "I've got a very serious face, Molly. People have commented on it." "That's why it's perfect. Children find serious clowns hilarious. Or terrifying. Either way, they'll be looking at you and not at me." Bill folded his arms. "And what are you doing while I'm being hilarious and terrifying?" "I'll be the bouncy castle attendant. At precisely three o'clock, I deflate the castle. Forty children are suddenly trapped under a giant rubber tent. Total chaos. Parents rush over. I roll the gift table straight out the side gate and into the van." "That seems cruel to the children." "They'll love it. It's like a nap with screaming." Bill considered this, then raised a finger. "What if the presents are rubbish? What if it's all socks?" "Nobody gives a seven-year-old socks. Except grandmothers, and I've accounted for grandmothers. We'll be filtering by shape. Rectangles are video games. Squishy ones go straight in the bin." "What about round ones?" Molly paused. She hadn't thought about round ones. "Round ones are a wildcard," she admitted. He nodded slowly, satisfied to have contributed. Then he pulled the clown costume out of the bag behind the seat and held it up. It was bright yellow, covered in polka dots, and cut for someone roughly half his height. "This is a child's costume." "It was a child's costume. It's a disguise now." "I won't be able to lift my arms." "Then don't lift your arms. Be a minimalist clown. Very modern. Very European." Bill stared at the tiny trousers for a long, sad moment. Across the road, a woman in a sun hat carried a giant stuffed giraffe up the Pembertons' driveway, and Molly's eyes lit up like a Christmas tree. "Bill," she whispered, gripping his arm. "That giraffe's worth forty quid if it's worth a penny." "How are we fitting a giraffe in the van?" Molly smiled the slow, terrible smile of a woman with a plan. "Neck first," she said.
created_at
"2026-09-23T17:56:56.223Z"
9/23/2026, 5:56:56 PM
model
"anthropic/claude-opus-5.5-20260921"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
4703
moderation_latency
(null)
generation_time
21301
tokens_prompt
75
tokens_completion
841
native_tokens_prompt
95
native_tokens_completion
1620
native_tokens_completion_images
(null)
native_tokens_reasoning
536
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
0.03278
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.20.0; linux; x64))"
http_referer
(null)
request_id
"req-1790186217-f1qsoRgNpSQejKwVYA8J"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790186217-ia5rvBKiwMIhfn8MoZaX"
upstream_id
"msg_011CfLoY27PFshDLcJBSWDSe"
provider_responses
0
endpoint_id
"3a2388bc-3740-4e64-a1f5-4e301726b6b9"
id
"msg_011CfLoY27PFshDLcJBSWDSe"
is_byok
false
latency
1395
model_permaslug
"anthropic/claude-opus-5.5-20260921"
provider_name
"Claude Platform on AWS"
status
200
total_cost
0.03278
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Claude Platform on AWS"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
97.63%
Matches word count
n/a
neededClean
false
words
507
27.27%
Dialogue to Total Word Ratio
Ratio: 55.10%, Deviation: 5.10%
neededClean
false
wordsTotal
510
wordsDialogue
281
62.4472%