NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-05-10-01KR8YF8NF9KXQVB4D0YGQ2NZ9
Run Details
Model:
Gemini 3.1 Flash Lite
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
34.5s
Cost:
$0.000899
Provider:
Google AI Studio
Evaluation summary
Score
Evaluator
Details Link
0.01%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
0.0049%
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 shoved the blueprint across the kitchen table, her finger tracing a jagged line around the backyard of the Miller estate. Bill stared blankly at the map, nursing a lukewarm coffee. Look, Bill, the structural integrity of the bouncy castle is the key. Its ventilation ducts are essentially a paved highway to the gift pile. Are you suggesting we enter through a bouncy castle, Molly? It seems structurally unsound. It is genius. The children are distracted by a magician with a severe gambling problem. We hit the presents during the communal singing of the Happy Birthday song. That song is the perfect acoustic cover for our tactical insertion. And how do we handle the primary opposition? The host parents. They hired a miniature pony. We can manipulate the pony to create a diversion. If I have to dress as a ride-along cowboy, I am taking an extra ten percent of the haul. You look great in a Stetson, Bill. Just focus on the logistics. We need a decoy gift, something heavy enough to look valuable but light enough to lug over a fence. I have a box filled with encyclopedias from nineteen-ninety-four. Perfect. Once we swap the boxes, we retreat via the vegetable patch. Are you sure the gate latch hasn’t been upgraded? I checked it twice. It’s held together by hope and structural rust. Any complications regarding the cake? The cake is our extraction timing anchor. Once the knife hits the frosting, the kids descend into a sugar-fueled frenzy. We vanish in that chaos. Do we have a contingency for the clown? I heard this clown is essentially a professional kickboxer moonlighting for extra cash. I am planning to lure him away with a trail of overpriced popcorn. He won't resist the salty snack. You really have thought of everything. I try my best. The party starts at three. We arrive at two-fifty, scale the trellis, and neutralize the perimeter. And the presents? We are talking about a mountain of Lego sets and high-end tech. The kids won't even notice the encyclopedias until Tuesday. What about the gift bags? The small ones at the exit? Leave them. Greed is the enemy of a clean getaway, Bill. Keep your eyes on the primary objective. You are right, I am getting ahead of myself. Focus, Bill. We are professionals. We are taking the presents because we need the liquidity, and because those children have too many toys anyway. It is practically a redistribution of wealth. A philanthropic heist, really. Exactly. Now, pass me the blueprints. I need to recalculate the drop-off point behind the oak tree. You want me to get the ropes ready? Yes, and get your fake mustache. The cowboy look isn't optional. Fine. But I am keeping the Stetson after this. Deal. Let’s go steal a birthday.
created_at
"2026-05-10T12:40:53.687Z"
5/10/2026, 12:40:53 PM
model
"google/gemini-3.1-flash-lite-20260507"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
484
moderation_latency
(null)
generation_time
34347
tokens_prompt
77
tokens_completion
680
native_tokens_prompt
67
native_tokens_completion
588
native_tokens_completion_images
0
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
"standard"
usage
0.00089875
router
(null)
provider_responses
0
endpoint_id
"237af107-d2e9-46c7-b146-c74d75980dee"
id
"1XwAasuRNoSsz7IPqqiR8Qk"
is_byok
false
latency
484
model_permaslug
"google/gemini-3.1-flash-lite-20260507"
provider_name
"Google AI Studio"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.1; linux; x64))"
http_referer
(null)
request_id
"req-1778416853-4TkinPYVBBDsaJUm8RIn"
session_id
(null)
api_type
"completions"
id
"gen-1778416853-vpk7EcsLDaS6ZACMRU4R"
upstream_id
"1XwAasuRNoSsz7IPqqiR8Qk"
total_cost
0.00089875
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Google AI Studio"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
0.01%
Matches word count
n/a
neededClean
false
words
469
0.00%
Dialogue to Total Word Ratio
Ratio: 0.00%, Deviation: 70.00%
neededClean
false
wordsTotal
476
wordsDialogue
0
0.0049%