NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYCR1VF3YC3K5AGPDJ2ESHK7
Run Details
Model:
Claude Opus 5 (Reasoning, Low)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
23.4s
Cost:
$0.031365
Provider:
Amazon Bedrock
Evaluation summary
Score
Evaluator
Details Link
27.17%
Matches word count
(details)
99.48%
Dialogue to Total Word Ratio
(details)
63.3254%
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 blueprints were drawn on the back of a placemat from the Golden Wok, which Bill considered a triumph of improvisation and Molly considered a health hazard. Sweet-and-sour sauce had bled through the paper, turning the west lawn into a swamp and the bouncy castle into what appeared to be a large orange tumor. Molly tapped the tumor with a chewed pen. "Explain this again." "Inflatable castle. Turrets, moat, the works. Rented from Party Palace for a hundred and forty quid a day, which frankly is highway robbery." "You'd know." "I would." Bill sat back, pleased. "Now. Presents get stacked here, on the trestle table, under the banner. Nobody opens them till cake. That's our window." They were in the back of a transit van parked outside number forty-two Bellweather Crescent, which belonged to a woman named Deborah who was, according to the neighbourhood's very loud grapevine, throwing her daughter Poppy a seventh birthday party with a magician, a petting zoo, and a chocolate fountain. Molly had scouted it twice. The chocolate fountain was the only part she resented having to leave behind. She studied the placemat with the grim patience of a woman who had once done six weeks in Holloway because Bill got peckish during a jewellery job and left a half-eaten Cornish pasty on a display case. "Guest list?" "Twenty-two children. Nine adults, six of whom will be drinking prosecco out of paper cups by eleven-fifteen, so really we're looking at three functional adults, and one of those is the magician, and magicians are cowards." "That's a stereotype." "It's a pattern I've observed." Molly circled the side gate. The side gate was the plan. The side gate was always the plan; in her professional experience, ninety percent of British domestic security was a latch and a hopeful attitude. She'd go over at 12:40, when the magician began the finale and every child in the garden turned to face the little folding table like sunflowers. Bill would bring the van around. Ninety seconds, in and out, and by one o'clock the presents would be in a lock-up in Tottenham while twenty-two children learned an important lesson about the fundamental cruelty of the universe. Bill was quiet for a moment, which was unusual and therefore ominous. "It's just," he said. "They're kids." "They're consumers." "Poppy's turning seven." "Poppy," said Molly, "is getting an iPad, a Nintendo, and something called a Hatchimal, which retails at forty-nine ninety-nine and screams. I'm doing her mother a favour." "You've got no soul." "I've got a mortgage. Same thing, worse interest." He grumbled, but he folded the placemat into his jacket, and Molly knew that was that, because Bill had never in eleven years talked himself out of anything except a diet. Outside, faintly, a bicycle bell rang. A dog barked. Somewhere over the fence, an inflatable castle inflated with a long asthmatic wheeze, rising up like a monument to everything Molly had never had. She checked her watch. "Balaclavas?" "Left them at Mum's." "Bill." "I brought party hats," he said, brightening. "Camouflage." Molly closed her eyes. It was, infuriatingly, better than balaclavas.
created_at
"2026-07-25T13:39:51.403Z"
7/25/2026, 1:39:51 PM
model
"anthropic/claude-opus-5-20260723"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3565
moderation_latency
(null)
generation_time
23321
tokens_prompt
75
tokens_completion
863
native_tokens_prompt
93
native_tokens_completion
1236
native_tokens_completion_images
(null)
native_tokens_reasoning
58
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
0.031365
router
(null)
provider_responses
0
endpoint_id
"76cb4608-f48c-483d-8da8-9957fb44244e"
id
"msg_011CdNsPzn5PYLH4zMZyVcVT"
is_byok
false
latency
1354
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-1784986791-ZJqEOtVwCUizWhCr4ncK"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1784986791-j1ILnRR919hK0S63b7tU"
upstream_id
"msg_011CdNsPzn5PYLH4zMZyVcVT"
total_cost
0.031365
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
27.17%
Matches word count
n/a
neededClean
false
words
519
99.48%
Dialogue to Total Word Ratio
Ratio: 29.68%, Deviation: 0.32%
neededClean
false
wordsTotal
529
wordsDialogue
157
63.3254%