NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-20-01KHYAVE1X5DKAJ89AHYH87WXW
Run Details
Model:
Gemma 3 12B
Test:
Dialogue tags
Scenario:
Write 200 words with 90% dialogue
Duration:
4.1s
Cost:
$0.000182
Provider:
Cloudflare
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
99.78%
Dialogue to Total Word Ratio
(details)
49.8896%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 200 word scene that has 90% 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
“Right, so, bouncy castle first?” Molly asked, tapping a crayon on the blueprint of the garden. Bill squinted. “Bouncy castle? Molly, we’re after the presents, not a sugar rush. Think, woman, think! Security.” “The bouncy castle *is* security, Bill! Distraction! Kids bouncing, parents distracted, we swoop in like… like fluffy, present-snatching eagles!” “Eagles don’t steal presents, Molly. They steal fish. And fluffy eagles are an oxymoron.” “Details, details! Look, there’s a gap in the hedge near the prize table. We go in, grab the loot, out before anyone notices a thing.” “And what about the clown? He looks… intense.” “Intense clowns are easily bribed. I’ve got a bag of jelly beans. Works every time.” “Jelly beans? You’re bribing a professional entertainer with *jelly beans*?” “Hey, they’re gourmet jelly beans! Lychee, mango, passionfruit… it’s a whole experience.” “Molly, this is madness. Utter, sugar-fueled madness.” “Exactly! Now, are you in or are you going to let a clown stand between us and a mountain of toys?”
created_at
"2026-02-20T20:11:25.641Z"
2/20/2026, 8:11:25 PM
model
"google/gemma-3-12b-it"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
205
moderation_latency
(null)
generation_time
3947
tokens_prompt
93
tokens_completion
300
native_tokens_prompt
75
native_tokens_completion
278
native_tokens_completion_images
(null)
native_tokens_reasoning
0
native_tokens_cached
0
num_media_prompt
(null)
num_input_audio_prompt
(null)
num_media_completion
0
num_search_results
(null)
origin
"https://ncbench.com/"
is_byok
false
finish_reason
"stop"
native_finish_reason
"stop"
usage
0.00018193
router
(null)
provider_responses
0
endpoint_id
"08ebad49-bed1-4149-850c-d24e3c1d2084"
id
"id-1771618285756"
is_byok
false
latency
110
model_permaslug
"google/gemma-3-12b-it"
provider_name
"Cloudflare"
status
200
api_type
"completions"
id
"gen-1771618285-iIs7L1ilKrqBq36SeLnM"
upstream_id
"id-1771618285756"
total_cost
0.00018193
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Cloudflare"
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
166
99.78%
Dialogue to Total Word Ratio
Ratio: 92.17%, Deviation: 2.17%
neededClean
false
wordsTotal
166
wordsDialogue
153
49.8896%