NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1K9MD2X027Q219CT4AQQNYG
Run Details
Model:
DeepSeek V4 Pro 0813 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
1m 51s
Cost:
$0.049295
Provider:
CoreWeave
Evaluation summary
Score
Evaluator
Details Link
99.92%
Matches word count
(details)
100.00%
Dialogue to Total Word Ratio
(details)
99.9595%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 50% 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 crouched behind a bouncy castle shaped like a dragon, her knees popping loud enough to startle a nearby pony handler. She clutched a clipboard and a pair of walkie-talkies, one still in its packaging. The backyard was a festival of primary colors, sugar, and shrieking six-year-olds. Bill slid next to her wearing a stolen catering vest and a paper crown. "How many presents?" Bill asked. "Thirty-eight, including the unicorn-wrapped box that's definitely a tablet." "We take the tablet first." "We take all of them. That's the plan." She unfolded a hand-drawn map of the gift table, which looked less like a heist blueprint and more like a toddler's crayon drawing of a traffic accident. Bill pointed at a lopsided X near the piñata. "What's this?" "The drop zone for the decoy gifts." "Decoy gifts?" "Empty boxes filled with rocks and old birthday cards." Bill frowned, and his paper crown tilted. "The kids will notice." "Not if we swap them during the piñata. Everyone watches the blindfolded kid swing at a unicorn." He nodded slowly, then stopped. "What about the clown?" "There are two clowns." "I'm saying we become the clowns." Molly stared at him. The catering vest suddenly seemed less like a disguise and more like a cry for help. "You want us to put on face paint and floppy shoes?" "Backup clowns. No one questions backup clowns." "Everyone questions backup clowns." "Not if we bring balloon animals." She considered this. A child ran past with a juice box, screaming with either joy or existential dread. The pony handler gave them a suspicious look. "Fine. But you handle the balloon animals." "Deal. I'll make a snake. Kids love snakes." "Kids are terrified of snakes." "Then they won't get in our way." Molly sighed and folded the map into a tiny square. "We leave the cake." "Obviously." "And we don't touch the piñata candy." Bill looked wounded. "What kind of monsters do you think we are?" "The kind stealing presents from a six-year-old." "Fair point." Bill adjusted the paper crown and peeked around the bouncy castle. A magician had started his act on the patio, and a dozen children sat cross-legged, hypnotized by a rabbit that looked deeply unimpressed. The gift table stood twenty feet away under a balloon arch shaped like a rainbow. Molly tapped her walkie-talkie. "Are we using code names?" Bill asked. "No." "I'll be Badger." "You'll be Bill." "Badger is more professional." Molly closed her eyes. "Fine, Badger. What's your brilliant plan for the bouncy castle?" Bill grinned. "We deflate it." Molly frowned. "That's just sad." "Sad is a distraction." "Fine. We'll deflate the bouncy castle." "And I'll steal the pony." Molly stared at him for a long moment. "We are not stealing a pony." Bill raised his hands. "I didn't say we'd keep it. I said we'd steal it." Molly rubbed her temples. "Why would we steal a pony?" "To carry the presents." "Ponies can't carry thirty-eight presents." "Then we steal two ponies." Molly said, "Now you're thinking."
created_at
"2026-09-03T09:29:37.13Z"
9/3/2026, 9:29:37 AM
model
"deepseek/deepseek-v4-pro-20260813"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
626
moderation_latency
(null)
generation_time
110971
tokens_prompt
75
tokens_completion
10426
native_tokens_prompt
146
native_tokens_completion
12400
native_tokens_completion_images
(null)
native_tokens_reasoning
9664
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
"stop"
service_tier
(null)
usage
0.04929526
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788427777-qOPVpjmWkUP4xTcsczKa"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788427777-nGObGBwo4iC2zox1Ycct"
upstream_id
"chatcmpl-5390daca-659d-48d7-8046-ca22009e62ea"
provider_responses
0
endpoint_id
"15560c07-2edc-40a6-b5ee-30819e66728a"
id
"chatcmpl-5390daca-659d-48d7-8046-ca22009e62ea"
is_byok
false
latency
182
model_permaslug
"deepseek/deepseek-v4-pro-20260813"
provider_name
"CoreWeave"
status
200
total_cost
0.04929526
cache_discount
(null)
upstream_inference_cost
0
provider_name
"CoreWeave"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
99.92%
Matches word count
n/a
neededClean
false
words
503
100.00%
Dialogue to Total Word Ratio
Ratio: 50.00%, Deviation: 0.00%
neededClean
false
wordsTotal
514
wordsDialogue
257
99.9595%