NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-20-01KHYAS5FBE14RJWHXT5FYH50B
Run Details
Model:
Gemma 3 4B
Test:
Dialogue tags
Scenario:
Write 200 words with 50% dialogue
Duration:
3.2s
Cost:
$0.000025
Provider:
DeepInfra
Evaluation summary
Score
Evaluator
Details Link
0.03%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
0.0152%
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
“Right, so, Operation Sugar Rush. We’re in, we grab everything shiny and bouncy, and we’re out. Simple,” Molly said, meticulously arranging miniature plastic dinosaurs on the table. “Simple? Molly, we’re infiltrating a five-year-old’s birthday party. ‘Simple’ doesn’t even cover the level of parental vigilance we’re likely to encounter,” Bill grumbled, chewing on the end of a pencil. “Are you *sure* about the dinosaurs? They look intensely judgmental.” “They’re camouflage, Bill! Little plastic reptiles blend in. Besides, think of the haul – twenty-seven Lego sets, a remote-controlled monster truck, a glittery unicorn…” “A glittery unicorn? Seriously? You’re going to risk a prolonged custody battle for a glittery unicorn?” “It’s fabulous! Look, the point is, we need a distraction. I was thinking a perfectly timed interpretive dance to ‘Baby Shark’?” “Absolutely not. We’re going in quiet. Stealth. Remember the briefing?” “Briefing? Bill, this is a birthday party, not a military operation. Just…promise me you won’t ask for a cupcake.” “No promises,” Bill replied, a mischievous glint in his eye.
created_at
"2026-02-20T20:10:11.318Z"
2/20/2026, 8:10:11 PM
model
"google/gemma-3-4b-it"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
187
moderation_latency
(null)
generation_time
3056
tokens_prompt
93
tokens_completion
316
native_tokens_prompt
75
native_tokens_completion
272
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
0.00002476
router
(null)
provider_responses
0
endpoint_id
"d3de3bd4-81bc-48fb-924f-2a87b2a36e75"
id
"chatcmpl-RQUPfOrxszoRfjW0EW7fBDKh"
is_byok
false
latency
68
model_permaslug
"google/gemma-3-4b-it"
provider_name
"DeepInfra"
status
200
api_type
"completions"
id
"gen-1771618211-E3c0QLTPyly0PN5blIJu"
upstream_id
"chatcmpl-RQUPfOrxszoRfjW0EW7fBDKh"
total_cost
0.00002476
cache_discount
(null)
upstream_inference_cost
0
provider_name
"DeepInfra"
Evaluation details
Result
Evaluator
Details
Meta Data
0.03%
Matches word count
n/a
neededClean
false
words
170
0.00%
Dialogue to Total Word Ratio
Ratio: 84.12%, Deviation: 34.12%
neededClean
false
wordsTotal
170
wordsDialogue
143
0.0152%