NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYCRFF9689TEZH32JZC1FA9G
Run Details
Model:
Claude Opus 5 (Reasoning, Low)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
20.8s
Cost:
$0.029290
Provider:
Amazon Bedrock
Evaluation summary
Score
Evaluator
Details Link
99.74%
Matches word count
(details)
83.65%
Dialogue to Total Word Ratio
(details)
91.6958%
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 was drawn on the back of a laminated menu from the Golden Wok, which meant that the ballroom of the Pemberton Country Club was rendered in ballpoint pen directly over a photograph of sweet-and-sour pork. Molly had circled the buffet table twice and drawn an arrow to it labeled TARGET ZONE. Bill kept staring at the pork. They were parked outside the club in a hatchback that smelled of wet dog, though neither of them had ever owned a dog. Molly had the seat reclined and a bag of gummy worms balanced on her sternum. Bill had a clipboard, because Bill believed in clipboards the way other men believed in God. "Nine years old," Molly said. "Only child. Two sets of divorced grandparents competing for affection. Bill, this is the richest vein in the county." "It's a birthday party." "It's an unguarded vault with a bouncy castle." Bill wrote something on the clipboard, then crossed it out. Through the window he could see the caterers wheeling in a cake shaped like a dinosaur, and beyond them, stacked in a glittering ziggurat against the far wall, the presents. Dozens of them. Some as large as microwaves. One appeared to be a bicycle wearing a bow. "There'll be parents," he said. "There'll be *fathers*," Molly corrected, sitting up. "Do you know what fathers do at children's birthday parties? They stand near the door and look at their phones and pray for death. By two o'clock they will have the situational awareness of a houseplant." "And the mothers?" "The mothers are the problem. The mothers see everything. That's why we're going in during the magician." Bill looked at her. Molly smiled the smile she reserved for moments when she had already won an argument he hadn't started yet. "Kevin the Astonishing," she said. "Forty-five minutes. Doves. Every adult in that room will be filming their kid's face for a group chat. Nobody films the corner where the loot is." The plan, as she laid it out, involved a laundry cart, two aprons stolen from a hotel in Doncaster, and a sign reading GIFT SORTING STATION — DO NOT DISTURB, which Molly had printed at a library in Comic Sans because, she explained, nothing looks less like crime than Comic Sans. Bill's role was to wheel the cart. Molly's role was to talk, which was always Molly's role, and which she had elevated into something close to art. "And if a child sees us?" "Children are the safest witnesses in the world. Tell them you're Santa's auditor." "It's June." "Tell them Santa's had a difficult fiscal year." Bill exhaled through his nose and made a final note on the clipboard, which read: *tell Molly no*, underlined twice, and which he already knew he would never say. Outside, a clown arrived on a unicycle and immediately fell off it. Molly took this as an omen, and refused to specify which kind. She handed him an apron. He put it on. "Forty-five minutes," she said. "Try to look employed."
created_at
"2026-07-25T13:47:17.679Z"
7/25/2026, 1:47:17 PM
model
"anthropic/claude-opus-5-20260723"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3916
moderation_latency
(null)
generation_time
20710
tokens_prompt
75
tokens_completion
809
native_tokens_prompt
93
native_tokens_completion
1153
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
0.02929
router
(null)
provider_responses
0
endpoint_id
"76cb4608-f48c-483d-8da8-9957fb44244e"
id
"msg_011CdNsxtVE3FN75YGr7ouo8"
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-1784987237-O9uKfWqmZ7C3vVGPawKJ"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1784987237-LpBQODfDN9xEW1TsIeWu"
upstream_id
"msg_011CdNsxtVE3FN75YGr7ouo8"
total_cost
0.02929
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
99.74%
Matches word count
n/a
neededClean
false
words
504
83.65%
Dialogue to Total Word Ratio
Ratio: 31.89%, Deviation: 1.89%
neededClean
false
wordsTotal
508
wordsDialogue
162
91.6958%