NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-05-29-01KSSJEQ9DSRR3CYA3FYG7VA73
Run Details
Model:
Claude Opus 4.8 (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
20.8s
Cost:
$0.029765
Provider:
Anthropic
Evaluation summary
Score
Evaluator
Details Link
98.71%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
49.3566%
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 blueprint across the table, smoothing it flat with both hands. "Okay, Bill, focus. The Hendersons' backyard. The presents are stacked here, on the gift table, right next to the bouncy castle." Bill leaned over the paper, squinting. "Why is there a clown drawn in the corner?" "That's not a clown. That's me. I'm wearing a disguise." "You drew yourself as a clown." "It's a strategic disguise, Bill. Nobody suspects the clown." "Everybody suspects the clown. Clowns are inherently suspicious." Molly slapped the table. "Can we please stay on target? The party starts at two. Cake at two-thirty. That's our window." "During cake?" Bill scratched his head. "That seems cruel. Stealing presents from a child while she's eating cake." "That's the genius of it. Sugar makes them sluggish. Distracted. A room full of frosted toddlers is the perfect cover." "What's the kid's name again?" "Tabitha. She's turning six." Bill winced. "Six. That's a tough age, Molly. You really wanna be the reason little Tabitha cries herself to sleep?" "She's rich, Bill. Look at this house. She's got a pony. A six-year-old does not need a pony and forty presents." "How do you know there's forty?" "I counted the RSVPs. I hacked the Evite." Bill stared at her. "You hacked a children's birthday invitation." "It wasn't hard. The password was 'Tabitha123.'" "That's a terrible password." "She's six, Bill, what do you want from her?" Bill sighed and tapped the blueprint. "Alright, alright. So I go in as what? The magician?" "No, you're the entertainment's assistant. You wheel in the supply cart, we load the presents under the tablecloth, and we walk out." "The supply cart." Bill nodded slowly. "And the actual magician?" "There is no magician. We are the magician." "So when the kids ask for a magic trick—" "You do a magic trick, Bill!" "I can't do magic trick!" "Pull a coin out of an ear or something! Improvise!" Bill folded his arms. "Last time I improvised, we got chased through a farmers market by a man with a baguette." "That baguette situation was not my fault." "You told him you were the cheese inspector!" "And he believed me!" Molly threw up her hands. "That's called confidence, Bill. That's the difference between a professional and a guy who panics." Bill went quiet, then pointed at the blueprint again. "What about the parents?" "Distracted. They'll all be filming Tabitha blow out candles. Phones up, backs turned. We've got ninety seconds." "And the dog?" Molly froze. "What dog?" "They have a golden retriever. I saw it on the Evite photos. Big fella. Name's Biscuit." "...Biscuit wasn't in my plan." Bill smiled, victorious. "See? This is why we case the joint properly. We need to neutralize Biscuit." "We're not hurting a dog, Bill." "I'm not hurting him! I'm bringing him a steak. Befriend the asset." Molly considered this, then nodded. "Okay. That's actually smart." "I have my moments." Bill rolled up the blueprint. "So. Clown suit's in the car?" "Clown suit's in the car." "God help us."
created_at
"2026-05-29T09:53:40.15Z"
5/29/2026, 9:53:40 AM
model
"anthropic/claude-4.8-opus-20260528"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
997
moderation_latency
(null)
generation_time
20731
tokens_prompt
77
tokens_completion
832
native_tokens_prompt
93
native_tokens_completion
1172
native_tokens_completion_images
(null)
native_tokens_reasoning
0
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
0.029765
router
(null)
provider_responses
0
endpoint_id
"b0d3beaa-5ff7-4666-a749-b1bb2b14c916"
id
"msg_01677vNCSzXTG6zBDUpe6Uo3"
is_byok
false
latency
997
model_permaslug
"anthropic/claude-4.8-opus-20260528"
provider_name
"Anthropic"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.16.0; linux; x64))"
http_referer
(null)
request_id
"req-1780048420-bQtOZF86AOr7mcbV3SUg"
session_id
(null)
api_type
"completions"
id
"gen-1780048420-wW8NNzaxGRlJ8QM0koNY"
upstream_id
"msg_01677vNCSzXTG6zBDUpe6Uo3"
total_cost
0.029765
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Anthropic"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
98.71%
Matches word count
n/a
neededClean
false
words
506
0.00%
Dialogue to Total Word Ratio
Ratio: 85.27%, Deviation: 15.27%
neededClean
false
wordsTotal
509
wordsDialogue
434
49.3566%