NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYCR5X89T7CNKS3QM24MB1PZ
Run Details
Model:
Claude Opus 5 (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
21.1s
Cost:
$0.030415
Provider:
Amazon Bedrock
Evaluation summary
Score
Evaluator
Details Link
86.38%
Matches word count
(details)
83.50%
Dialogue to Total Word Ratio
(details)
84.9390%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 30% 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 blueprint spread across the hood of Bill's Corolla was, in fact, a placemat from the Cracker Barrel on Route 9, but Molly had drawn on it in red marker and that made it official. She had labeled the bounce house PRIMARY OBSTACLE and the piñata SECONDARY OBSTACLE and, in the corner, a small circle with the word GRANDMA underlined three times. Bill leaned in, chewing a toothpick he'd been working since Tuesday. He smelled faintly of WD-40 and regret. "Explain the grandma." "Ninety-one years old. Sits by the gift table the entire party. Doesn't eat, doesn't move, doesn't blink. Judy DeMarco. She's the wall." "So we go around the wall." "You don't go around Judy DeMarco. You go through her, or you go home." The party was tomorrow at two, in the backyard of a colonial on Sycamore, in honor of a child named Tyler who was turning seven and who, according to the invitation Molly had lifted from a neighbor's recycling bin, had asked for "Legos and also money." Forty-one guests. Molly had counted the RSVPs on the mother's Facebook post, then counted them again, because the number of presents determined whether this was a job or a hobby. Bill's contribution to the planning process had so far consisted of a rented costume, still in its bag, which he now produced with the reverence of a man unveiling a sword. "What is that." "It's a llama." "It's a llama." "Petting zoo's booked. I called. I said I was the mother and I wanted to add a llama, and they said they don't have a llama, and I said that's fine, I've got one." Molly closed her eyes. Somewhere down the block a leaf blower started up, mourning. The plan, as it eventually crystallized over a shared bag of gas station pretzels, went like this. Bill, in the llama, would enter through the side gate at 2:40, the traditional hour of maximum sugar and minimum supervision. He would then perform what Molly kept referring to as "a distraction of unusual quality," a phrase Bill did not care for and could not get her to define. Meanwhile Molly, dressed as a woman who belonged there — sunglasses, iced coffee, an expression of profound tiredness — would drift to the gift table, engage Judy DeMarco in a discussion of her hip, and begin relocating packages into an oversized tote bag reading LIVE LAUGH LOVE. "How many trips?" "Three. Four if the Peterson kid brings the wagon." "And the getaway?" "Minivan. Nobody looks twice at a minivan leaving a birthday party. It's the most invisible vehicle in America. You could transport a body in a minivan." "That's dark, Molly." "That's freedom, Bill." They stood a moment in the golden light of the cul-de-sac, contemplating the years of their lives that had led here. Bill folded the placemat carefully and slid it into his jacket. Molly finished the pretzels. Above them, a single balloon, escaped from some earlier and unrelated celebration, drifted westward over the rooftops like an omen neither of them was qualified to read.
created_at
"2026-07-25T13:42:04.305Z"
7/25/2026, 1:42:04 PM
model
"anthropic/claude-opus-5-20260723"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3972
moderation_latency
(null)
generation_time
20955
tokens_prompt
75
tokens_completion
826
native_tokens_prompt
93
native_tokens_completion
1198
native_tokens_completion_images
(null)
native_tokens_reasoning
72
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
0.030415
router
(null)
provider_responses
0
endpoint_id
"76cb4608-f48c-483d-8da8-9957fb44244e"
id
"msg_011CdNsZnjK7hmFNRkgHT7bW"
is_byok
false
latency
1405
model_permaslug
"anthropic/claude-opus-5-20260723"
provider_name
"Amazon Bedrock"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.18.0; linux; x64))"
http_referer
(null)
request_id
"req-1784986924-zyB7zH2qAUjrnnu480LA"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1784986924-tkvjpfzBooDzVB9DpnqD"
upstream_id
"msg_011CdNsZnjK7hmFNRkgHT7bW"
total_cost
0.030415
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Amazon Bedrock"
response_cache_source_id
(null)
data_region
"global"
Evaluation details
Result
Evaluator
Details
Meta Data
86.38%
Matches word count
n/a
neededClean
false
words
511
83.50%
Dialogue to Total Word Ratio
Ratio: 28.10%, Deviation: 1.90%
neededClean
false
wordsTotal
516
wordsDialogue
145
84.9390%