NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M37PKDTD5RY2SR7HVD2KXCTY
Run Details
Model:
Claude Opus 5.5 (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
17.2s
Cost:
$0.023920
Provider:
Claude Platform on AWS
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
54.86%
Dialogue to Total Word Ratio
(details)
77.4282%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 70% 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 hood of the van and pinned one corner down with a juice box. "Right. Sunday, two o'clock. Tyler Henderson turns seven. Forty guests, one bouncy castle, and a present table the size of a small boat." Bill squinted at the drawing. "Why is the house shaped like a potato?" "Because I drew it from memory while jogging past, Bill. Focus." "You don't jog." "I jogged that day. For the mission." She tapped a crayon circle. "This is the gift table. It's by the back door, which gives us a clean exit to the alley." "And how do we get in? Pose as caterers?" "Better." Molly reached into a duffel bag and pulled out a fluffy purple costume with a sagging unicorn head. "Entertainment." Bill stared at it for a long moment. "I'm not wearing that." "You're not. I am. You're wearing this." She produced a second costume: a rainbow wig, a red nose, and trousers so enormous they could have housed a family of four. "A clown? Molly, children can smell fear. They'll know I'm a fraud." "You are a fraud. That's the whole job." Bill held the trousers against his waist and sighed. "What do clowns even do?" "Balloon animals. I bought you a starter kit." She tossed him a bag of balloons. "Practice." Bill inflated one, twisted it three times, and held up a lumpy pink shape. "Dog." "That's a sausage." "It's a dachshund. They're basically sausages." He added a fourth twist and the balloon exploded with a crack that sent a pigeon fleeing from the roof. "Okay. Sick dachshund." Molly pressed on. "Here's the play. At two-fifteen, I lead the kids in the Unicorn Dance. Very distracting. Lots of spinning. Meanwhile, you start loading presents into your trousers." "Into my trousers?" "Why do you think they're so big? I measured. Those pockets hold eleven medium-sized gifts or one remote-control helicopter." Bill peered inside the waistband with new respect. "There's a shelf in here." "Adjustable. I'm not an amateur." "What about the parents? Parents at these things watch like hawks." "The parents will be at the drinks table by then. Mrs. Henderson makes a punch that's mostly rum. I checked her recycling bin." "You went through her bins?" "Research, Bill. Every great heist starts with garbage." He frowned at the potato-shaped house. "And the birthday boy? Tyler?" Molly's expression darkened. "Tyler is the wild card. According to my sources, he bit a magician last year." "Your sources?" "His cousin. She's nine. She wanted two dollars and a sticker." Bill swallowed. "He bit a magician?" "On the ankle. Right through the sock. So whatever you do, don't let him see the trousers moving." "Why would the trousers move?" "Hamster ball. It's on the wish list. Fully assembled, with the hamster already inside." Bill set down his balloons very carefully. "Molly, I think we need a plan B." She smiled and pulled a second crayon from her pocket. "Oh, Bill. This *is* plan B."
created_at
"2026-09-23T17:56:47.07Z"
9/23/2026, 5:56:47 PM
model
"anthropic/claude-opus-5.5-20260921"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
17121
moderation_latency
(null)
generation_time
17121
tokens_prompt
75
tokens_completion
791
native_tokens_prompt
95
native_tokens_completion
1177
native_tokens_completion_images
(null)
native_tokens_reasoning
70
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
0.02392
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.20.0; linux; x64))"
http_referer
(null)
request_id
"req-1790186207-IqI3jC6tMWYjCBbUTNJ3"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790186207-KC8dZuYOpNFeNwpEjiFf"
upstream_id
"msg_011CfLoXHECHThotQeSD9Dun"
provider_responses
0
endpoint_id
"3a2388bc-3740-4e64-a1f5-4e301726b6b9"
id
"msg_011CfLoXHECHThotQeSD9Dun"
is_byok
false
latency
1245
model_permaslug
"anthropic/claude-opus-5.5-20260921"
provider_name
"Claude Platform on AWS"
status
200
total_cost
0.02392
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
54.86%
Dialogue to Total Word Ratio
Ratio: 66.53%, Deviation: 3.47%
neededClean
false
wordsTotal
505
wordsDialogue
336
77.4282%