NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-10-08-01M4DT97KZZ550DQ4EC7J4ZZ0B
Run Details
Model:
Claude Haiku 5.5 (Reasoning, Max)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
3m 1s
Cost:
$0.023470
Provider:
Claude Platform on AWS
Evaluation summary
Score
Evaluator
Details Link
99.98%
Matches word count
(details)
96.11%
Dialogue to Total Word Ratio
(details)
98.0458%
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
The getaway van smelled like cheese puffs and bad decisions. Molly sat behind the wheel, squinting across the street at a backyard full of bobbing balloons and a banner that read HAPPY SEVENTH, MAX! Bill spread a map across the dashboard. It had been drawn on the back of a pizza box. A rainbow wig sat in his lap like a dozing cat. "Phase one," he whispered. "I go in as the party clown." "You're going in as a clown?" "I've been practicing. My clown name is Sprinkles." "That's the worst clown name I've ever heard." "Sprinkles is a very dangerous name, Molly. Nobody suspects a Sprinkles." "Fine. Phase two?" "We grab the gifts from the gazebo before the cake comes out. Then we deal with the bouncy castle." "What's wrong with the bouncy castle?" "Hazel." "Who's Hazel?" "Six years old. Juice box in one hand, whistle in the other. Her job is watching the presents." "Can't we bribe her?" "I already tried. I offered her a Rocket Pop." "What did she say?" "She said she's allergic to rockets." "Bill, we're robbing a six-year-old." "We're relieving her of responsibility. Children are very stressed." "And Grandma?" "Grandma Dot has a whistle too. We'll need earplugs." "Do we have earplugs?" "We have cotton balls. That's basically earplugs if you're brave." "We're going to need a better plan." "We have a better plan. We have the ice cream truck." Molly glanced at the van's side, where a cartoon cone grinned beside a painted cow. Not exactly subtle. "This isn't an ice cream truck, Bill. It's a van with a jingle button." "Exactly. Nobody suspects an ice cream truck. Everybody chases one." "Have you ever actually seen anyone chase one?" "I tested it outside the library." "And?" "Two toddlers and a librarian. Today we'll get a dozen kids. They'll sprint toward us, and we'll be gone before they realize we're not selling Rocket Pops." "So I'm the ice cream man?" "You're the ice cream man, Molly. You have the face for it. Trustworthy. A little sad." "Thanks. Do I get a hat?" "You get a cone hat." "And bring the wheelbarrow." "Why do we need a wheelbarrow?" "Some of those boxes are the size of a fridge." "What's in them, a refrigerator?" "Mostly dinosaurs. Also a remote-control shark." "I want the shark." "Nobody gets the shark, Molly. The shark is a menace." Molly sighed. "Do you feel bad about this?" "Not even a little. I'm a professional." "Bill, your lip is trembling." Bill dabbed white greasepaint onto his cheeks with a shaking finger. "It's the greasepaint, Molly. It's very stiff." "Showtime, Sprinkles." Molly hit the jingle button. The tune tinkled down the street, and within seconds a dozen kids dropped their juice boxes and sprinted toward the van. Hazel abandoned her post. Only Grandma Dot stayed put, whistle between her teeth, eyes narrowed. "Remind me again why we don't have normal jobs?" Molly asked. "Nobody hires a Sprinkles," Bill said, and slipped out of the van.
created_at
"2026-10-08T13:12:15.753Z"
10/8/2026, 1:12:15 PM
model
"anthropic/claude-haiku-5.5-20261007"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2335
moderation_latency
(null)
generation_time
180767
tokens_prompt
75
tokens_completion
7250
native_tokens_prompt
95
native_tokens_completion
46920
native_tokens_completion_images
(null)
native_tokens_reasoning
45717
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
0.0234695
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1791465135-g52uly5KlHy63IzKWael"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1791465135-kuyGI5J4ylHsiT7hBCZX"
upstream_id
"msg_011Cfppxd3Eh9tThdevPvHuh"
provider_responses
0
endpoint_id
"b6cdf493-3a61-441f-a134-e7944fddb980"
id
"msg_011Cfppxd3Eh9tThdevPvHuh"
is_byok
false
latency
742
model_permaslug
"anthropic/claude-haiku-5.5-20261007"
provider_name
"Claude Platform on AWS"
status
200
total_cost
0.0234695
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
99.98%
Matches word count
n/a
neededClean
false
words
502
96.11%
Dialogue to Total Word Ratio
Ratio: 70.89%, Deviation: 0.89%
neededClean
false
wordsTotal
505
wordsDialogue
358
98.0458%