NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYCRB5FS8S5WDYKSMTXRYWZH
Run Details
Model:
Claude Opus 5 (Reasoning, Low)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
20.3s
Cost:
$0.028440
Provider:
Amazon Bedrock
Evaluation summary
Score
Evaluator
Details Link
90.48%
Matches word count
(details)
0.11%
Dialogue to Total Word Ratio
(details)
45.2981%
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 napkin across the table and smoothed it flat with the side of her hand. "Right," she said. "Bouncy castle here. PiƱata here. Presents here, on the trestle table, under the banner that says Kai Is Six." "Kai," Bill said. "What kind of a name is Kai?" "The kind of name that comes with two sets of grandparents and a father in cryptocurrency. Focus." "I'm focused. I'm just saying, it feels a bit much, robbing a six-year-old." "He's not six. He's six *tomorrow*. Today he's five, and five-year-olds have no legal standing whatsoever." "That's not how the law works, Molly." "It's how *my* law works." She tapped the napkin. "Fourteen presents minimum. Probably twenty. There's a drone in there, Bill. A drone, a scooter, and something in a box the size of a fridge that the mother has been posting about for a fortnight." Bill leaned back and folded his arms. "And how do we get in? It's a garden party. Fences. Adults. Adults who have *phones*." "We go in as entertainment." "Absolutely not." "You'd make a lovely clown." "I've told you before, I don't do clowning. My uncle did clowning and it took the family thirty years to recover." "Then be a magician. Magicians get a table. Magicians get near the presents. Magicians carry a *big black case* nobody's allowed to look inside." Bill went quiet. It was the quiet of a man discovering that a terrible plan had a good bone in it. "How big a case?" "Big enough for a drone and a scooter, if you take the wheels off." "And the fridge-sized box?" "That's why I'm coming as your assistant. Assistants carry things. Nobody questions an assistant carrying something out to the van. It's practically the job description." "And what if a parent asks to see a trick?" "Do a trick." "I don't know any tricks, Molly." "You know the one with the coin." "That's not a trick, that's a coin going down my sleeve and staying there. I've lost eleven pounds fifty to that trick." "Then you'll do the one where the children close their eyes and count to a hundred while the magic happens." Bill stared at her. "That's not a trick either. That's just an instruction." "It's the greatest trick ever performed and it's never once failed. Children love counting. They'll be at forty-one before anyone realises the fridge box has gone." Molly folded the napkin into quarters. "Van's parked on Lyndhurst Road, engine running, boot open. Ninety seconds from the gate." "And the cake?" "What about the cake?" "Are we taking the cake?" "Bill." "It's a big cake. She posted about the cake as well." Molly looked at him for a long moment, then picked up her pen and added a small rectangle to the napkin, just beside the trestle table. "Fine," she said. "But you're carrying it, and if you drop it, I'm telling everyone you were the clown."
created_at
"2026-07-25T13:44:56.578Z"
7/25/2026, 1:44:56 PM
model
"anthropic/claude-opus-5-20260723"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3471
moderation_latency
(null)
generation_time
20183
tokens_prompt
75
tokens_completion
770
native_tokens_prompt
93
native_tokens_completion
1119
native_tokens_completion_images
(null)
native_tokens_reasoning
49
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
0.02844
router
(null)
provider_responses
0
endpoint_id
"76cb4608-f48c-483d-8da8-9957fb44244e"
id
"msg_011CdNsnVYpAubxXREBfB4Vp"
is_byok
false
latency
1226
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-1784987096-EVAjH6PvzebbSk1OeMGT"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1784987096-NlQnSznqX5GpzY9JX99K"
upstream_id
"msg_011CdNsnVYpAubxXREBfB4Vp"
total_cost
0.02844
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
90.48%
Matches word count
n/a
neededClean
false
words
490
0.11%
Dialogue to Total Word Ratio
Ratio: 81.65%, Deviation: 11.65%
neededClean
false
wordsTotal
496
wordsDialogue
405
45.2981%