NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-05-10-01KR90TEX0HCP3GW85C5FW4734
Run Details
Model:
Gemma 4 31B (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
5m 51s
Cost:
$0.003914
Provider:
DeepInfra
Evaluation summary
Score
Evaluator
Details Link
3.62%
Matches word count
(details)
100.00%
Dialogue to Total Word Ratio
(details)
51.8117%
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 slapped a crayon-drawn map onto the stained coffee table with a dramatic flourish, her eyes wide with intensity. The room smelled of old coffee and desperation. The map was mostly scribbles. "Listen up, Bill. This is the target: The Miller residence. Six years old. High energy. High security." Bill scratched his chin, squinting at the jagged lines and smudge marks. "Is that a dinosaur or a sofa?" "It's the sectional, you idiot," she hissed, pointing aggressively at a blob of brown wax. "We slip in during the cake cutting." Bill sighed heavily, leaning back in his creaky wooden chair that groaned under his weight. He looked at the ceiling, wondering where it all went wrong. "Why are we stealing from kids, Molly? This feels remarkably low." "Low? It's a 'Tech-Themed' party," Molly explained, her voice dropping to a conspiratorial whisper. "The den is full of tablets. It's a goldmine." Bill stared at the peeling beige paint of the ceiling, wondering if he could find a more moral partner in crime. "What's the disguise?" "We're 'Entertainment Specialists,'" she said, pulling two neon-colored costumes from a dusty, moth-eaten garment bag. The fabrics were shimmering and loud. "I've got the face paint and foam fingers. You're the balloon artist." "I don't actually know how to make balloons," Bill protested, poking a limp piece of pink latex with a look of pure disgust. "You don't need to make them, you just need to look busy while I bag the loot." Molly pointed to a red X on the map, her finger trembling with sheer excitement. The laundry room offered a quick escape to the alley. "The presents are staged here. We strike fast, and we exit through the laundry room." "What if the kids see us?" Bill asked, imagining a swarm of sticky-handed toddlers. "They're six, Bill. They have the attention span of goldfish. As long as there's frosting, we're invisible." Bill looked at the giant foam finger, wondering if he could still get his old job at the warehouse. "Is this really necessary?" "Professionalism is everything," Molly snapped, straightening her tactical vest over her glittery party dress. "Do you want to be a specialist?" "I just want a job where I don't have to wear a giant yellow glove." "Tough. Now, let's review the extraction. Once the presents are in the duffel, we hit the fence by midnight." "Who is the fence for Legos?" Bill asked, genuinely curious about the logistics of plastic bricks. "The underground brick market is booming, Bill. Now, put on the glove." Bill slowly slid the foam finger onto his hand, looking completely defeated. "I feel ridiculous." "You look like a professional," Molly corrected, checking her stopwatch with a smug smirk. "And remember, you're from 'Party-Palooza Inc.'" "Do we have a license for this?" "We have a rusty van and a dream. Now move!"
created_at
"2026-05-10T13:21:57.671Z"
5/10/2026, 1:21:57 PM
model
"google/gemma-4-31b-it-20260402"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
263
moderation_latency
(null)
generation_time
349981
tokens_prompt
77
tokens_completion
9523
native_tokens_prompt
82
native_tokens_completion
10272
native_tokens_completion_images
(null)
native_tokens_reasoning
8780
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
0.00391402
router
(null)
provider_responses
0
endpoint_id
"4149ab03-3b69-45a0-b94c-6c253b8641a1"
id
"chatcmpl-Rm9mMBb4XEGOxqKJqpQCim0T"
is_byok
false
latency
53
model_permaslug
"google/gemma-4-31b-it-20260402"
provider_name
"DeepInfra"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.1; linux; x64))"
http_referer
(null)
request_id
"req-1778419318-PMfdXpoI4jWwnWVwZWI6"
session_id
(null)
api_type
"completions"
id
"gen-1778419318-CdCy51WsNJJozSNQ2EmX"
upstream_id
"chatcmpl-Rm9mMBb4XEGOxqKJqpQCim0T"
total_cost
0.00391402
cache_discount
(null)
upstream_inference_cost
0
provider_name
"DeepInfra"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
3.62%
Matches word count
n/a
neededClean
false
words
476
100.00%
Dialogue to Total Word Ratio
Ratio: 50.00%, Deviation: 0.00%
neededClean
false
wordsTotal
482
wordsDialogue
241
51.8117%