NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-10-08-01M4DAKXQX7XBES2D4H8135M1H
Run Details
Model:
Claude Haiku 5.5 (Adaptive)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
10.8s
Cost:
$0.001095
Provider:
Claude Platform on AWS
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
66.42%
Dialogue to Total Word Ratio
(details)
33.2114%
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 crouched behind the hedge, peering through binoculars at the glowing bounce castle in the Hendersons' backyard. Bill tugged at his ski mask, which had started itching the moment he put it on. "Remind me again why we're robbing a seven-year-old's birthday party," Bill whispered. "Because the Hendersons are rich, and rich people buy expensive presents," Molly said. "Also, the loan shark's deadline was Tuesday, and I'd rather not sleep with one eye open." "Fair. What's the plan?" "Phase one: we slip through the side gate while the clown is doing balloon animals. Phase two: we grab every wrapped box with a bow on it. Phase three: we vanish like ghosts." "Phase three sounds vague." "Phase three is the best part. Trust me." Bill squinted. "There's a man in a giant cupcake costume by the gate." "That's Gary. He's our inside man." "Gary's been in a cupcake costume since noon." "He's committed to the bit." Molly tossed him a walkie-talkie. "Gary, sitrep." Static crackled. "Sitrep. I'm dressed as a cupcake, Molly. There are forty kids here and one of them just bit my frosting." "Is the present table unguarded?" "Define unguarded. Grandma's sitting on it. She's knitting." Bill groaned. "We're not sneaking past a grandma." "Grandmas are the easiest. They fall asleep in five minutes." "Says who?" "Says the Hendersons' Christmas party last year, when my uncle swapped seats with her and she didn't notice for an hour." Bill hesitated. "Okay, but if she wakes up yelling about her yarn, I'm leaving you behind." "Deal. Now stop talking. We've got a window." They crept across the lawn, bent low, until a sprinkler sprayed them both. Bill nearly yelped. "Quiet!" hissed Molly. "You're quiet! You're dripping on the tulips!" They reached the table. Molly lifted a box wrapped in glittery paper. It rattled. "What's in this one?" Bill asked. "No idea. Could be a puppy." She shook it. "Actually, that sounds like Lego." "Lego is heavy, Molly." "Lego is valuable, Bill. Grab the big ones." Grandma snored. Bill froze. "Did she just—" "Keep moving." They stuffed presents into a laundry bag as fast as they could. Behind them, a small voice piped up. "Mommy, the bad people are taking my presents!" Molly turned slowly. A five-year-old in a pirate hat pointed a plastic sword at them. "Ahoy," Bill said weakly. "Give them back," the pirate demanded. Molly sighed and handed over the Lego. "Fine. But you didn't see us." "I saw you." "Then you didn't see us with conviction."
created_at
"2026-10-08T08:38:28.868Z"
10/8/2026, 8:38:28 AM
model
"anthropic/claude-haiku-5.5-20261007"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
10814
moderation_latency
(null)
generation_time
10814
tokens_prompt
75
tokens_completion
649
native_tokens_prompt
95
native_tokens_completion
2171
native_tokens_completion_images
(null)
native_tokens_reasoning
1164
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
0.001095
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1791448708-xGvuWkCA84OdpjD3clnn"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1791448708-sgg3jbWESB5LlQvnt1vD"
upstream_id
"msg_011CfpU5iokk7W5DogH9JuXm"
provider_responses
0
endpoint_id
"b6cdf493-3a61-441f-a134-e7944fddb980"
id
"msg_011CfpU5iokk7W5DogH9JuXm"
is_byok
false
latency
735
model_permaslug
"anthropic/claude-haiku-5.5-20261007"
provider_name
"Claude Platform on AWS"
status
200
total_cost
0.001095
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
0.00%
Matches word count
n/a
neededClean
false
words
418
66.42%
Dialogue to Total Word Ratio
Ratio: 67.14%, Deviation: 2.86%
neededClean
false
wordsTotal
423
wordsDialogue
284
33.2114%