NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M37PP8JT747VNKC3VJB5S3BJ
Run Details
Model:
Claude Opus 5.5 (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
30.7s
Cost:
$0.051080
Provider:
Claude Platform on AWS
Evaluation summary
Score
Evaluator
Details Link
93.65%
Matches word count
(details)
8.64%
Dialogue to Total Word Ratio
(details)
51.1456%
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 unrolled the blueprint across the dashboard of the van, pinning one corner under a thermos and the other under Bill's sandwich. "That's not a blueprint," Bill said. "That's a birthday invitation. It has a unicorn on it." "The unicorn is a landmark. Focus." Molly tapped the glittery card. "Saturday, two o'clock, 14 Maple Drive. Tyler is turning seven. Forty guests confirmed. That's forty presents, Bill. Minimum." Bill chewed thoughtfully, which for Bill was a visible, strenuous process. "What if some kids bring cards instead?" "Then we steal the cards." "What if the cards just say 'Happy birthday, love Grandma'?" "Then we'll have Grandma's love, which is more than you've got now." The van was parked two houses down, painted with the words BALLOON ANIMALS BY BARTHOLOMEW. Molly had chosen the disguise because nobody ever questioned a clown. Bill had chosen the name Bartholomew because it was, in his words, "extremely trustworthy." "Here's the plan," Molly said. "You go in as the entertainment. While the kids are mesmerized by your balloon animals, I slip into the gift room, load everything into laundry bags, and we're gone before the cake." Bill raised his hand. "We're in a van. You don't have to raise your hand." "I can't make balloon animals." Molly stared at him. "You told me you could make balloon animals." "I said I could make a balloon snake." "What's a balloon snake?" "It's a balloon, Molly. You blow it up and you believe." Molly closed her eyes, the way she did whenever she remembered she'd been kicked out of three separate crime syndicates for her choice of partners. When she opened them, Bill was holding up a limp pink tube with the pride of a new father. "His name is Gerald," Bill said. "Fine. Gerald is the distraction. Tell them Gerald is shy." "Seven-year-olds are ruthless. They'll eat me alive." "They're children." "My nephew's party had a bouncy castle. Two kids went in. One came out wearing the other one's shoes." Molly consulted her notes, written in crayon because the van's previous owner had stolen all the pens. The van's previous owner was also Molly. "Other complications. PiƱata at three." "So?" "So forty children with a stick. Stay clear of the backyard. And the mother, Linda, runs the neighborhood watch Facebook group. Four thousand members." Bill went pale. "The Maple Drive Moms." "You know them?" "They found my bike in a day. I didn't even report it stolen. I stole it." Molly folded the invitation and tucked it into her jacket. "In and out, Bill. Nobody gets hurt. Nobody gets glitter on them." "Somebody always gets glitter on them." She sighed, because he was right, and he was so rarely right that it felt like an omen. Down the street, a minivan pulled up and a small boy climbed out, dragging a present the size of a refrigerator. Bill whispered, "Molly. I think it's a bike." "Start the van, Bartholomew."
created_at
"2026-09-23T17:58:20.011Z"
9/23/2026, 5:58:20 PM
model
"anthropic/claude-opus-5.5-20260921"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3788
moderation_latency
(null)
generation_time
30651
tokens_prompt
75
tokens_completion
1139
native_tokens_prompt
95
native_tokens_completion
2535
native_tokens_completion_images
(null)
native_tokens_reasoning
1448
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
0.05108
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.20.0; linux; x64))"
http_referer
(null)
request_id
"req-1790186300-JlapiHMxp5SluH8CPsUG"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790186300-JN53A9B29saD5O4AmrlZ"
upstream_id
"msg_011CfLoe8aF5fVNCNtqzt7tq"
provider_responses
0
endpoint_id
"3a2388bc-3740-4e64-a1f5-4e301726b6b9"
id
"msg_011CfLoe8aF5fVNCNtqzt7tq"
is_byok
false
latency
1148
model_permaslug
"anthropic/claude-opus-5.5-20260921"
provider_name
"Claude Platform on AWS"
status
200
total_cost
0.05108
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
93.65%
Matches word count
n/a
neededClean
false
words
491
8.64%
Dialogue to Total Word Ratio
Ratio: 57.00%, Deviation: 7.00%
neededClean
false
wordsTotal
493
wordsDialogue
281
51.1456%