NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-06-01-01KT1H2JWZR9CYTGDJSTMH88SP
Run Details
Model:
Claude Opus 4.8 (Reasoning, Low)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
21.5s
Cost:
$0.029640
Provider:
Anthropic
Evaluation summary
Score
Evaluator
Details Link
98.71%
Matches word count
(details)
1.97%
Dialogue to Total Word Ratio
(details)
50.3390%
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 out the wrinkles with the reverence of a surgeon. "Behold," she said. "The Henderson residence. Site of one Timmy Henderson's seventh birthday party." Bill leaned in, squinting. "Why is the blueprint drawn in crayon?" "Because I had to blend in at the reconnaissance. You try casing a joint surrounded by toddlers without arts and crafts." "Fair." He tapped the paper. "So what's the plan?" "We infiltrate during the magician's act. Nobody watches the present pile when there's a man pulling rabbits out of a hat." "I hate magicians." "Everyone hates magicians, Bill. That's why it's the perfect cover." She slid a photograph toward him. "This is our target. Mountain of gift-wrapped boxes, estimated value—" "Forty dollars?" "Don't be cynical. There could be a Nintendo in there." Bill picked up the photo and frowned. "Molly, that's a clown." "That's our second obstacle. Bingo the Clown. Hired security, basically. He never leaves the snack table." "So we neutralize the clown." "We do not neutralize the clown." Molly snatched the photo back. "We distract him. With cake." "What if he doesn't want cake?" "Everyone wants cake, Bill. That's basic human psychology." "Clowns aren't human. They're a separate, terrifying species." Molly pinched the bridge of her nose. "Can we please focus? Now, I'll go in disguised as Aunt Linda. Nobody knows what Aunt Linda looks like because she lives in Phoenix." "And me?" "You're the entertainment. Balloon animals." Bill stood up so fast his chair toppled. "Absolutely not. I don't do balloons." "You did three years for stealing balloons!" "That was helium tanks, and you know it." He crossed his arms. "There's a difference between grand theft inflation and twisting a poodle for a four-year-old." "It's a disguise!" "It's humiliating!" Molly took a deep breath. "Bill. Sweetheart. Light of my criminal career. We have rent due Friday and a fence who wants his cut. We are stealing presents from a child." "When you say it out loud it sounds bad." "It is bad. We're thieves. That's the job." She set a hand on his shoulder. "Now. Can you make a balloon dog or not?" He sighed. "I can make a balloon snake." "A snake is just a balloon you gave up on." "It's abstract art." "Fine. You're an abstract balloon artist." She rolled up the blueprint. "We go in at two o'clock. Magician starts the rabbit trick, clown gets distracted by cake, we grab the presents, out the back by the bouncy castle." "And if something goes wrong?" "What could possibly go wrong at a seven-year-old's birthday party?" Bill stared at her. "You said that before the petting zoo job." "The petting zoo job was a learning experience." "I still have the scar." "Goats are unpredictable, Bill. Nobody could have foreseen that." She extended her hand. "So. Are we doing this?" He looked at the crayon blueprint, then at her, then sighed and shook her hand. "For the record," he said, "if there's a piñata, I'm taking it." "Obviously. The candy's untraceable."
created_at
"2026-06-01T12:03:29.317Z"
6/1/2026, 12:03:29 PM
model
"anthropic/claude-4.8-opus-20260528"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1131
moderation_latency
(null)
generation_time
21482
tokens_prompt
77
tokens_completion
829
native_tokens_prompt
93
native_tokens_completion
1167
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
0.02964
router
(null)
provider_responses
0
endpoint_id
"b0d3beaa-5ff7-4666-a749-b1bb2b14c916"
id
"msg_011qatH9v2HvPTUUmWqAq2BF"
is_byok
false
latency
1131
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-1780315409-uf0tFXHXSWUypX4nx8RF"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1780315409-aZlwnE3yRiIrnSatJmUj"
upstream_id
"msg_011qatH9v2HvPTUUmWqAq2BF"
total_cost
0.02964
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
1.97%
Dialogue to Total Word Ratio
Ratio: 78.86%, Deviation: 8.86%
neededClean
false
wordsTotal
511
wordsDialogue
403
50.3390%