NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-29-01M3P2D5MVA2V18DSMBJG2W66T
Run Details
Model:
Claude Sonnet 5.5 (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
20.4s
Cost:
$0.025000
Provider:
Claude Platform on AWS
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.47%
Dialogue to Total Word Ratio
(details)
0.2361%
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 hood of the van, pinning its corners with a juice box and three mini-donuts. "Okay, Bill. Saturday, two o'clock, Bounce House Palace. The birthday boy is Tyler, age seven, and according to his mother's very public social media, he's getting forty-one presents." "Forty-one," Bill whispered. "That's more than I got in my entire childhood." "That's why you became a thief." "That's why I became a thief," he agreed, wiping a tear. "So what's the plan?" "Simple. You go in as the clown. I go in as the party planner. We gather the gifts, load the van, and leave before cake." "Nobody leaves before cake, Molly. It looks suspicious." "Fine. We leave after cake but before the piñata." "Why before the piñata?" "Because last time, Bill, you got emotional and hit a child." "He was holding the candy hostage! And he had a very aggressive stance for a seven-year-old." Molly sighed. "Just stay in character. What's your clown name?" "Bonko." "Bonko's taken. The real clown is booked for that party." "Then I'll be Sir Honks-a-Lot." "That's worse. That's a lawsuit." "What about the real clown?" "Handled. He's locked in a porta-potty behind the Dairy Barn with a sandwich and a podcast." Bill nodded approvingly. "Humane. I like that." "We have standards." She tapped the blueprint. "The gifts are stacked by the trampoline wall. Your job is distraction. Make balloon animals and keep the parents busy." "I can't make balloon animals." "Bill." "I can make one shape. It's sort of a sad snake." "Then make lots of sad snakes." "Do parents like those?" "No. But they'll be too confused to look at me." Bill frowned at the map. "What about security?" "One guy. Gary. He guards the shoe cubbies." "Shoe cubbies?" "Everyone removes their shoes for the bounce house. So do we. Silent as ninjas." "Ninjas in socks." "Exactly. Did you bring socks?" Bill lifted his pant leg, revealing a sock covered in cartoon dinosaurs. "Stegosaurus." "Perfect. You'll blend right in." "What if the presents are just clothes?" "Nobody gives a seven-year-old clothes at a party, Bill. Only grandmothers." "Will there be grandmothers?" "There are always grandmothers. Avoid the one with the casserole. She has seen things." Bill shuddered. "And the getaway?" Molly grinned and pulled a tiny pink scooter from the back of the van. "Gary can't chase what he can't catch." "Molly, that's a child's scooter." "It's a decoy. We're taking the van." "Then why did you bring the scooter?" She paused. "I got excited at the store." Bill nodded solemnly and picked up a mini-donut. "Forty-one presents." "Forty-one presents," she echoed, clinking her juice box against his donut.
created_at
"2026-09-29T07:52:27.04Z"
9/29/2026, 7:52:27 AM
model
"anthropic/claude-sonnet-5.5-20260928"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3480
moderation_latency
(null)
generation_time
20426
tokens_prompt
75
tokens_completion
1040
native_tokens_prompt
95
native_tokens_completion
2481
native_tokens_completion_images
(null)
native_tokens_reasoning
1416
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
0.025
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1790668347-uNvFEz8STtfJ14Ja6cx9"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790668347-OhbTFQ0DPExTSVIhLteG"
upstream_id
"msg_011CfXNHtmrZVGNSNXPCq2oA"
provider_responses
0
endpoint_id
"99aaad94-923b-4fc1-b763-271ed5486f7a"
id
"msg_011CfXNHtmrZVGNSNXPCq2oA"
is_byok
false
latency
648
model_permaslug
"anthropic/claude-sonnet-5.5-20260928"
provider_name
"Claude Platform on AWS"
status
200
total_cost
0.025
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
445
0.47%
Dialogue to Total Word Ratio
Ratio: 80.35%, Deviation: 10.35%
neededClean
false
wordsTotal
458
wordsDialogue
368
0.2361%