NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M37P9WA1TH6ZSWJW372ESACJ
Run Details
Model:
Claude Opus 5.5 (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
31.9s
Cost:
$0.058280
Provider:
Claude Platform on AWS
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
98.15%
Dialogue to Total Word Ratio
(details)
99.0735%
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
Molly spread the blueprints across the hood of the van, pinning one corner with a thermos and the other with Bill's elbow. "That's not a blueprint," Bill said. "That's a birthday invitation." "It's a map, Bill. It has a map on the back." She tapped a crayon arrow. "See? 'Park by the big tree, come around to the bouncy castle.' They've told us exactly where the security is." Bill squinted. Someone had drawn a dinosaur wearing a party hat next to the words TIMMY IS SEVEN!!! The exclamation points had been gone over so many times the paper had torn. Twenty years in the business, Molly thought. Banks, jewelers, one regrettable museum. And now this. But the jewelry trade had dried up, the museums had installed lasers, and a seven-year-old in the suburbs had forty-three RSVPs. She had counted. Forty-three guests, each carrying a wrapped box. That was a haul. "Walk me through it again," Bill said, which was what he said whenever he hadn't been listening the first time. She sighed and unfolded the second sheet, a real diagram, drawn by her, in pen, like a professional. The backyard. The gift table. The fence gate. A small circle labeled CLOWN? "Why's the clown got a question mark?" "Because I don't know if there's a clown. Clowns are wild cards. Clowns have seen things. Clowns fight back." Bill nodded gravely. He had been chased by a clown once, in Reno, and did not like to discuss it. The plan was simple, which was how Molly liked them. At two-fifteen, Bill would enter through the side gate in disguise and announce himself as the entertainment. While forty-three children and thirty exhausted parents watched him, Molly would slip behind the hydrangeas, load the gifts into a laundry cart, and roll them out to the van. Six minutes, tops. Nobody watched the gift table during the entertainment. The disguise sat folded on the passenger seat. It was purple and furry and had cost Molly eighty dollars at a costume shop that also sold, inexplicably, bait. "It smells like worms," Bill added. "I have questions about the entertainment part," Bill said. "You're a magician." "I don't know any magic." "You can pick a lock in four seconds. Pick a lock in front of them. Kids love that." "That's not magic, Molly. That's evidence." She considered this. It was, annoyingly, a fair point. She looked at the purple suit, then at Bill, then at the crayon dinosaur, which seemed to be judging her. "Fine," she said. "Can you juggle?" "I can juggle three things." "What three things?" Bill thought for a long moment. "Oranges. Or a phone, a wallet, and keys, but only if they belong to someone else." Molly folded the diagram and tucked it into her jacket. Somewhere across town, a mother was hanging streamers and a boy named Timmy was vibrating with anticipation, and none of them had any idea what was coming. "Buy oranges," she said. "And practice. Timmy deserves a show."
created_at
"2026-09-23T17:51:34.23Z"
9/23/2026, 5:51:34 PM
model
"anthropic/claude-opus-5.5-20260921"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
4788
moderation_latency
(null)
generation_time
31883
tokens_prompt
75
tokens_completion
1127
native_tokens_prompt
95
native_tokens_completion
2895
native_tokens_completion_images
(null)
native_tokens_reasoning
1848
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
0.05828
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.20.0; linux; x64))"
http_referer
(null)
request_id
"req-1790185894-kMpglhQ8sucZj4PuZD9J"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790185894-rjuqUZs4qGBnxpdXOiEc"
upstream_id
"msg_011CfLo8E2pT9PtUXYWZYGpx"
provider_responses
0
endpoint_id
"3a2388bc-3740-4e64-a1f5-4e301726b6b9"
id
"msg_011CfLo8E2pT9PtUXYWZYGpx"
is_byok
false
latency
1181
model_permaslug
"anthropic/claude-opus-5.5-20260921"
provider_name
"Claude Platform on AWS"
status
200
total_cost
0.05828
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
100.00%
Matches word count
n/a
neededClean
false
words
501
98.15%
Dialogue to Total Word Ratio
Ratio: 29.39%, Deviation: 0.61%
neededClean
false
wordsTotal
507
wordsDialogue
149
99.0735%