NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-20-01KHXPK00FMSK0CP28Q8AXE3FD
Run Details
Model:
Mistral Medium 3.1
Test:
Dialogue tags
Scenario:
Write 200 words with 10% dialogue
Duration:
8.6s
Cost:
$0.000628
Provider:
Mistral
Evaluation summary
Score
Evaluator
Details Link
95.99%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
47.9934%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 200 word scene that has 10% 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
The rainbow-balloon arch over the bounce castle sagged slightly in the afternoon heat, its vibrant colors clashing with the grim determination in Molly’s eyes as she adjusted her oversized clown wig. Beside her, Bill fiddled with a party popper, his fingers twitching like he was defusing a bomb instead of a child’s toy. A cluster of sugar-high six-year-olds shrieked in the distance, their tiny voices rising above the tinny rendition of *"Baby Shark"* blaring from a Bluetooth speaker. Molly nudged Bill with her elbow. “You sure the gift table’s unguarded?” Bill squinted through his fake glasses—glasses that, upon closer inspection, had one lens smudged with what looked suspiciously like jelly. “Only if you don’t count the mom with the death grip on her coffee cup. She’s got the stance of a woman who’s seen *Taken* one too many times.” “Then we distract her.” Molly produced a whoopee cushion from her fanny pack with the gravitas of a surgeon unsheathing a scalpel. “You cause a diversion. I’ll liberate the LOL Dolls.” Bill groaned. “Last time you said ‘diversion,’ I ended up in a bouncy castle with a broken tailbone.” “And yet, here you are.” Molly grinned, adjusting her fake nose. “Professional to the end.”
created_at
"2026-02-20T14:17:17.595Z"
2/20/2026, 2:17:17 PM
model
"mistralai/mistral-medium-3.1"
3/1/2001, 12:00:00 AM
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
375
moderation_latency
(null)
generation_time
8564
tokens_prompt
77
tokens_completion
333
native_tokens_prompt
69
native_tokens_completion
300
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
0.0006276
router
(null)
provider_responses
0
endpoint_id
"13c5002b-8c5b-490c-8a5f-52ec0242804f"
id
"cb372d0b8bf24eb485d8ad01e4f69bbe"
is_byok
false
latency
375
model_permaslug
"mistralai/mistral-medium-3.1"
3/1/2001, 12:00:00 AM
provider_name
"Mistral"
status
200
api_type
"completions"
id
"gen-1771597037-5RTvCylPS9wDy4DzUlJ0"
upstream_id
"cb372d0b8bf24eb485d8ad01e4f69bbe"
total_cost
0.0006276
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Mistral"
Evaluation details
Result
Evaluator
Details
Meta Data
95.99%
Matches word count
n/a
neededClean
false
words
208
0.00%
Dialogue to Total Word Ratio
Ratio: 36.06%, Deviation: 26.06%
neededClean
false
wordsTotal
208
wordsDialogue
75
47.9934%