NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-10-08-01M4DZYEWCE98QQKXAMQQRMMBT
Run Details
Model:
Claude Haiku 5.5 (Reasoning, Max)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
3m 12s
Cost:
$0.025700
Provider:
Claude Platform on AWS
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
89.17%
Dialogue to Total Word Ratio
(details)
94.5861%
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
Bill squinted at the greasy pizza box balanced on the dashboard, where Molly had drawn the entire heist in red marker. "Why is there a bouncy castle in the middle of it?" "Because every kid under ten wants to be in it, and every parent is watching the magician." Molly tapped a crooked X by the hall doors. "Side door, two-fifteen. Pick the lock, grab the presents, and nobody looks up." "And if somebody does?" "Then it's Plan B." "What's Plan B?" "Plan A, but louder." "Where are the presents?" "Under the banner, next to the cake table. All forty-two gifts, mostly Lego. Bring the big sack, Bill. Lego is heavier than it looks." "How do we tell the good ones from the rubbish?" "We don't. We take the lot, Bill." "Even the jumpers?" "Especially the jumpers. Nobody deserves a jumper on their birthday." "What about the kids? There's a girl called Ruby who bit a policeman through his trousers." "Ruby's six. She'll be in the castle until she throws up, then in the loos, then asleep by the sausage rolls." "That's a lot of sugar for one little girl." "It's a children's party. Sugar is the whole point." "What if the kids cry?" "They'll cry for twenty minutes, then someone gives them a balloon dog and they forget their own names." Bill nodded slowly, then pointed through the windscreen. "And the getaway?" Across the road sat a rusty ice cream van, its speaker warbling a broken version of "Greensleeves." "Why's it playing Greensleeves?" "It's the only tune it's got." "We're escaping in an ice cream van?" "Nobody checks an ice cream van. People just wave at it." "That's two very different reasons to be nervous." "Stop being nervous, Bill. You're the one going in as the clown." Bill went very quiet. Reluctantly, he lifted a carrier bag and pulled out a rainbow wig, a squeaky red nose and a pair of enormous yellow shoes. "Molly, I have to tell you something about the clown." "Don't." "The real clown called in sick, so I rang the mum and said we were from an agency." "We?" "Me and my cousin Dennis." "Where's Dennis now?" Bill winced. "Dennis is the ice cream man." They both looked across the road. The van door swung open, and a large man in a paper hat waved at them cheerfully. "Oh, Bill." "He's very good at it." "He's terrifying at it." Bill pulled on the shoes. They squeaked. He squeezed the red nose experimentally, and it honked. Molly closed her eyes. "Right," she said. "Once you're inside, you do not, under any circumstances, juggle." "I can juggle." "You can juggle two satsumas and a sock." "That's still juggling." "And when we're out, it's four minutes, tops." "Four minutes?" "Six, if the cake's still on the table." "Why would the cake matter?" Molly grabbed the red marker and underlined the word CAKE on the pizza box, twice. "Because, Bill, I'm not leaving a party without a slice."
created_at
"2026-10-08T14:51:14.202Z"
10/8/2026, 2:51:14 PM
model
"anthropic/claude-haiku-5.5-20261007"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2397
moderation_latency
(null)
generation_time
191812
tokens_prompt
75
tokens_completion
6796
native_tokens_prompt
95
native_tokens_completion
51380
native_tokens_completion_images
(null)
native_tokens_reasoning
50274
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
0.0256995
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1791471074-gLJRrPzvPzUbT8OEY28C"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1791471074-NohHJOSeA45oqJWN1S7K"
upstream_id
"msg_011CfpxWNPdX5NZ3furakiHD"
provider_responses
0
endpoint_id
"b6cdf493-3a61-441f-a134-e7944fddb980"
id
"msg_011CfpxWNPdX5NZ3furakiHD"
is_byok
false
latency
729
model_permaslug
"anthropic/claude-haiku-5.5-20261007"
provider_name
"Claude Platform on AWS"
status
200
total_cost
0.0256995
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
100.00%
Matches word count
n/a
neededClean
false
words
500
89.17%
Dialogue to Total Word Ratio
Ratio: 71.51%, Deviation: 1.51%
neededClean
false
wordsTotal
502
wordsDialogue
359
94.5861%