NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-05-31-01KSYX8106SDZG1969AZ84WCP5
Run Details
Model:
Claude Opus 4.8 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
21.8s
Cost:
$0.028540
Provider:
Anthropic
Evaluation summary
Score
Evaluator
Details Link
86.38%
Matches word count
(details)
29.39%
Dialogue to Total Word Ratio
(details)
57.8829%
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 blueprints across the table, smoothing out the wrinkles with the flat of her hand. "Okay, so the perimeter is guarded by what intelligence suggests is a clown named Bozo," she said. Bill leaned in, squinting. "Armed?" "Balloon animals. But don't underestimate him. I've seen a man take a poodle to the eye and never fully recover." "Noted." Bill cracked his knuckles. "What's the layout?" "Bounce house here, snack table there, and the presents—" she tapped the center of the page, "—stacked on a folding table near the piñata." "A piñata?" Bill's face fell. "Molly, you know how I feel about piñatas." "Bill, that was years ago." "It exploded, Molly. Candy everywhere. I still find Smarties in my socks." "It was a Tuesday and you were emotionally fragile. We've all moved on." She slid a photograph across the table. "Here's our target. Birthday boy, age seven, goes by 'Tyler.'" Bill studied the image. "He looks smug." "He's seven." "Smug for seven." Molly ignored him. "We go in during the cake ceremony. Maximum distraction. Every adult will be filming on their phones, every kid will be screaming about frosting." "And the presents?" "We load them into the van disguised as a bouncy castle rental service." Bill nodded slowly, then frowned. "Wait. We're stealing presents from a child." "Brilliant deduction." "No, I mean—" he scratched his head, "—isn't that a little... low? Even for us?" "Bill, last month we robbed a retirement home's bingo fund." "That was different. Those were adults who made bad life choices. This is a kid." Molly sighed and pulled out a second photograph. "Look closer at the gift table." Bill leaned in. His eyes widened. "Is that—" "A limited edition gold-plated action figure. Worth forty grand. Some rich uncle gifted it, the fool." "Forty grand?" Bill whistled. "For a doll?" "Action figure. And yes. The kid won't even know what he's got. He'll probably try to eat it." Bill paced the room. "I don't know, Molly. The clown, the piñata, the moral ambiguity—" "Three hundred percent return on investment." "What's our exit?" "There it is." She grinned. "We blend in with the parents during the goodbye-bag distribution. Nobody questions two adults carrying loot if they're holding a goodie bag full of stickers." "And if Bozo gets suspicious?" "Then you do what you do best." Bill straightened up, suddenly proud. "Distract him with charm?" "No. Run. You run, Bill." "I can run." "You can run beautifully. Like a startled gazelle." He smiled despite himself. "Fine. I'm in. But I have one condition." "Name it." "We grab a slice of cake on the way out. I'm not robbing a party and leaving hungry. That's just bad manners." Molly considered this, then extended her hand. "Cake on the way out. Deal." They shook firmly. "One more thing," Bill added, glancing at the blueprints. "What flavor is the cake?" "Does it matter?" "Molly. It always matters."
created_at
"2026-05-31T11:38:27.423Z"
5/31/2026, 11:38:27 AM
model
"anthropic/claude-4.8-opus-20260528"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1142
moderation_latency
(null)
generation_time
21509
tokens_prompt
77
tokens_completion
809
native_tokens_prompt
93
native_tokens_completion
1123
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
0.02854
router
(null)
provider_responses
0
endpoint_id
"b0d3beaa-5ff7-4666-a749-b1bb2b14c916"
id
"msg_01YKsfwkmKcqmcCd8L5RtKKN"
is_byok
false
latency
1142
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-1780227507-JqHaXOBE0gbOohIW3Edk"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1780227507-7SFYhOm06PolIDFQ661j"
upstream_id
"msg_01YKsfwkmKcqmcCd8L5RtKKN"
total_cost
0.02854
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Anthropic"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
86.38%
Matches word count
n/a
neededClean
false
words
489
29.39%
Dialogue to Total Word Ratio
Ratio: 74.95%, Deviation: 4.95%
neededClean
false
wordsTotal
491
wordsDialogue
368
57.8829%