NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1KBC15RPCN467MVZKATC4D0
Run Details
Model:
DeepSeek V4 Flash 0731 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 200 words with 50% dialogue
Duration:
4.8s
Cost:
$0.000132
Provider:
Baidu
Evaluation summary
Score
Evaluator
Details Link
99.92%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
49.9595%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 200 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 flattened the crumpled invitation against the minivan’s dash. “We hit at three-fifteen, right when they’re doing the piñata. Zero adult supervision.” Bill squinted through the windshield at the pastel bouncy castle. “Three-fifteen? That’s when the magician does the balloon animals. The parents watch that. It’s a crowd-pleaser.” “Which is why we go for the Gift Table. One swift yank of the tablecloth, presents slide right into the duffel, we’re gone before the first dad yells ‘Ta-da!’” She mimed the magician’s flourish. Bill chewed a carrot stick. “What about the birthday kid?” “What about him?” “He’s, like, five. He’s gonna cry.” “He’ll get new presents. Probably better ones. Have you seen the wrapping on this stuff? It’s all dinosaurs and unicorns. That’s mid-tier retail, Bill. We’re not hitting a Gucci store, we’re hitting a yard sale with frosting.” “Still feels low.” Molly grabbed his wrist. “Low is a basement full of stolen LEGO sets. Low is selling a princess castle back to the mom who bought it. We’re heroes, Bill. We’re teaching emotional resilience through material loss.” Bill sighed. “Fine. But I’m keeping the unicorn. That one’s for me.” “Deal. Now wipe that guilt off your face. We’ve got a party to crash.”
created_at
"2026-09-03T09:59:59.943Z"
9/3/2026, 9:59:59 AM
model
"deepseek/deepseek-v4-flash-20260731"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
731
moderation_latency
(null)
generation_time
4596
tokens_prompt
75
tokens_completion
424
native_tokens_prompt
146
native_tokens_completion
399
native_tokens_completion_images
(null)
native_tokens_reasoning
70
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
0.00013216
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788429599-BVrNY0XLJQlx1nTsFe8p"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788429599-LL9avcgWHXKcR3ecjwTq"
upstream_id
"as-6nwiax06nz"
provider_responses
0
endpoint_id
"4d5286fc-8f37-40c2-a518-e1ae100d45fc"
is_byok
false
latency
74
model_permaslug
"deepseek/deepseek-v4-flash-20260731"
provider_name
"Wafer"
status
429
1
endpoint_id
"223c54e7-4db1-4098-980f-47741e7637ab"
id
"as-6nwiax06nz"
is_byok
false
latency
731
model_permaslug
"deepseek/deepseek-v4-flash-20260731"
provider_name
"Baidu"
status
200
total_cost
0.00013216
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Baidu"
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
203
0.00%
Dialogue to Total Word Ratio
Ratio: 83.01%, Deviation: 33.01%
neededClean
false
wordsTotal
206
wordsDialogue
171
49.9595%