NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1KA64QCGDHKRKBTDQ5VY6S5
Run Details
Model:
DeepSeek V4 Flash 0731 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
12.0s
Cost:
$0.000296
Provider:
Venice
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
0.0001%
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 squinted through the binoculars at the inflatable castle, then lowered them with a grimace. “Bill, I’m seeing a lot of juice boxes. That’s not intel, that’s a sugar rush waiting to happen.” Bill, who was lying on his stomach in the hedge, adjusted his ill-fitting balaclava. “Focus. The presents are under the tarp next to the bouncy castle. They’ve got a unicorn theme. It’s very secure.” “Is the unicorn armed?” Molly asked. “No, but the mom with the clipboard is. She’s got the organizational fury of a general,” Bill said. “We go in at 2:30, during the piñata. That’s when chaos peaks.” “You want to steal gifts from five-year-olds during a sugar rush? That’s not a heist, that’s a war crime,” Molly said. She picked a leaf out of her hair. “I need a better reason. What’s the haul?” Bill pulled a crumpled list from his pocket. “Twelve wrapped boxes. I did recon. Belt-tagged one as ‘From Grandma.’ It’s heavy. Could be a gaming console.” “Could be a casserole dish,” Molly countered. “Grandmas are treacherous.” “Okay, what’s your brilliant plan? Walk in with a party hat and a fake smile?” Molly grinned. “We don’t rob the kids. We rob the parents. You create a diversion—fake a twisted ankle right by the snack table. While everyone’s yelling ‘ice pack,’ I swap the gift tarp with my decoy tarp.” Bill blinked. “You made a decoy tarp.” “I made a decoy tarp,” she said proudly. “It’s printed with the same unicorns. But underneath? Empty boxes wrapped in yesterday’s newspaper.” “That’s … actually good,” Bill admitted. “But what if a kid sees you?” “I’ll tell them I’m the unicorn fairy. Kids love that,” Molly said. “And you’ll be crying like a baby, so no one’s looking at me.” “I don’t cry on command.” “I’ll step on your foot for real,” she offered. “Genuine tears.” Bill sighed, rubbing his ankle preemptively. “And when the parents open the decoy boxes at the end of the party? There’s going to be screaming. Possibly police.” Molly waved a hand. “By then we’re long gone. Also, I left a note inside each decoy box that says ‘The real gift was the friends you made along the way.’ They’ll think it’s a life lesson.” “You’re a monster.” “I’m a professional,” Molly corrected. She checked her watch. “Piñata’s in twenty. You ready to take one for the team?” Bill pulled off his balaclava and used it to wipe his face. “Fine. But if I see a single disappointed kid, I’m telling them your address.” Molly laughed. “Deal. But if the grandma’s box really is a casserole, you’re making me dinner.” She crawled out of the hedge, brushed off her knees, and winked. “Let’s go steal some joy.” Bill followed, muttering, “I need a better agent.”
created_at
"2026-09-03T09:39:18.394Z"
9/3/2026, 9:39:18 AM
model
"deepseek/deepseek-v4-flash-20260731"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
700
moderation_latency
(null)
generation_time
11370
tokens_prompt
75
tokens_completion
809
native_tokens_prompt
146
native_tokens_completion
772
native_tokens_completion_images
(null)
native_tokens_reasoning
50
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
0.00029575
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788428358-7sl3wHj5yM0cWo6Wtsnm"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788428358-UX9e4UiCkjwTB6G0APCW"
upstream_id
"chatcmpl-3e7e8f32ddc2488596f742303306f5bf"
provider_responses
0
endpoint_id
"2da47185-782a-4f49-bb59-aa7cf7f026ec"
is_byok
false
latency
143
model_permaslug
"deepseek/deepseek-v4-flash-20260731"
provider_name
"Fireworks"
status
429
1
endpoint_id
"223c54e7-4db1-4098-980f-47741e7637ab"
is_byok
false
latency
449
model_permaslug
"deepseek/deepseek-v4-flash-20260731"
provider_name
"Baidu"
status
429
2
endpoint_id
"f45eb809-95d6-4a0b-8b13-7f56a7b86375"
id
"chatcmpl-3e7e8f32ddc2488596f742303306f5bf"
is_byok
false
latency
700
model_permaslug
"deepseek/deepseek-v4-flash-20260731"
provider_name
"Venice"
status
200
total_cost
0.00029575
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Venice"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
466
0.00%
Dialogue to Total Word Ratio
Ratio: 76.17%, Deviation: 26.17%
neededClean
false
wordsTotal
470
wordsDialogue
358
0.0001%