NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-05-10-01KR8W1V7MX892B1WFR7WR3HKQ
Run Details
Model:
Gemma 4 26B
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
19.8s
Cost:
$0.000348
Provider:
Venice
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
0.0000%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 70% 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 leaned over the blueprint, which was actually just a hand-drawn map of a suburban backyard on a napkin. Bill, stop chewing on your pencil. You’re making me nervous. I’m not nervous, Molly. I’m calibrating. It’s a tactical snack. It’s a No. 2 Ticonderoga, and it’s distracting. Now, listen. The target is the bouncy castle. That’s where the heavy lifting happens. The bouncy castle? That’s amateur hour. We should go for the gift table. It’s stationary. It’s predictable. It’s high-yield. The gift table is a death trap. It’s right next to the juice box station. You know how sticky those things are. One slip and we’re stuck to the tablecloth like flies on jam. We need mobility. We need the bouncy castle. Bill, the bouncy castle is filled with screaming toddlers. It’s a chaotic environment. Exactly! Chaos is our camouflage! If a kid sees us, we just pretend we’re part of the inflatable obstacle course. We’ll be 'The Birthday Ninjas.' You are not wearing a ninja suit, Bill. We agreed on tactical black. Ninja suits are breathable. Tactical black is just heavy polyester that makes you sweat like a marathon runner in a sauna. Plus, ninjas have better stealth ratings. Stealth ratings? Did you read a comic book again? It was a technical manual! Focus. The plan is this: I slide under the perimeter fence, you provide lookout from behind the hydrangea bushes. When the cake is being cut, that’s our window. The cake window is risky. What if there’s a piñata? A piñata is a diversionary tactic. It draws all eyes to a single point of impact. It’s classic misdirection. It’s a paper donkey filled with Skittles, Bill. It’s not a distraction, it’s a snack break. To a professional, everything is a distraction. Fine. If we go for the bouncy castle, how do we extract the loot? We can’t exactly walk out with a giant teddy bear tucked under our arms. We use the balloon animal technique. The what? I’ll twist some long balloons into makeshift carrying straps. It’s genius. Nobody suspects a guy with a balloon poodle. You are going to stand in a backyard, surrounded by children, and pretend to be a balloon artist to cover up a grand larceny? It’s called deep cover. It’s called a felony waiting to happen. Is it still a felony if the victims are under the age of seven? Yes, Bill! Yes, it is! Do you think they have those little party favors? The plastic whistles? Those could be useful for signaling. Get out. I’m just saying, a whistle is a versatile tool. Get out of my sight and go check the getaway van. Fine, but if I see a clown, I’m calling it off. Clowns are unpredictable. They’re the wild cards of the birthday circuit.
created_at
"2026-05-10T11:58:36.796Z"
5/10/2026, 11:58:36 AM
model
"google/gemma-4-26b-a4b-it-20260403"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
550
moderation_latency
(null)
generation_time
19651
tokens_prompt
77
tokens_completion
733
native_tokens_prompt
79
native_tokens_completion
671
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)
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
0.000348337
router
(null)
provider_responses
0
endpoint_id
"2e9daaa4-87bf-49f7-a44a-2c1d617c2c51"
id
"chatcmpl-a9765708ab22ba83"
is_byok
false
latency
550
model_permaslug
"google/gemma-4-26b-a4b-it-20260403"
provider_name
"Venice"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.1; linux; x64))"
http_referer
(null)
request_id
"req-1778414316-izgc4nMalWOE9Yt8YBmK"
session_id
(null)
api_type
"completions"
id
"gen-1778414316-2VFzqPH4fQNUUhF4uSwv"
upstream_id
"chatcmpl-a9765708ab22ba83"
total_cost
0.000348337
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Venice"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
465
0.00%
Dialogue to Total Word Ratio
Ratio: 0.64%, Deviation: 69.36%
neededClean
false
wordsTotal
467
wordsDialogue
3
0.0000%