NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-10-05-01M46BNWD2398ERJS8QKK9SGEG
Run Details
Model:
GPT-6.1 Sol (Reasoning, Medium)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
28.9s
Cost:
$0.008972
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
99.74%
Matches word count
(details)
10.70%
Dialogue to Total Word Ratio
(details)
55.2197%
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 spread a crayon drawing across the diner table. It showed a house, a smiling sun, and something that might have been a dog or an exploding potato. Bill leaned over it, careful not to put his elbow in the syrup. “That’s the target,” Molly said. “Birthday party. Saturday. Twenty children, thirty presents. We go in empty-handed and come out looking like Christmas had a custody dispute.” “Why is the house purple?” “Because I borrowed the crayons from a witness.” “Is the witness reliable?” “She’s six. She described the security system as ‘a mean goose,’ which I think we should take seriously.” Bill turned the drawing sideways. He had once cracked a hotel safe with a stethoscope, three paper clips, and remarkable patience. Unfortunately, it had belonged to the hotel’s magician, and everything inside was scarves. He still carried one for emergencies. “What’s our cover?” he asked. “Parents? I can do disappointed father. I’ve got the eyebrows.” “No. Entertainers. You’re a magician, I’m your assistant. We arrive with an enormous box, announce a disappearing-present trick, and leave before anyone asks how it works.” “That’s theft with a bow tie.” “All theft should have a bow tie. It reassures people.” “What if they want me to perform?” “Pull something out of a hat.” “Last time I tried that, the rabbit bit the vicar.” “That was a ferret, Bill.” “It was dark inside the hat.” Molly produced an invitation from her coat. Glitter drifted onto the table between them. Bill brushed a speck from his sleeve, missed it, and tried again. The invitation smelled faintly of strawberry lip balm and looked entirely beyond his professional experience. “Whose birthday is it?” “Oliver’s. Age seven. Theme: pirates.” “Pirates? Molly, that’s competition.” “It’s children wearing cardboard hats.” “And adults pretending piracy is charming. We spend years developing a brand, and suddenly everybody gets an eyepatch.” “We’ll fit right in. You already limp.” “That’s an occupational injury.” “You fell off a coin-operated pony.” “It bucked.” For a moment they studied the heap of presents pictured beside the purple house. Bill reached for the invitation, then stopped to clean syrup from his fingers. Molly watched him read the bottom line twice. She put down her pencil. “What?” she said. “It says, ‘Please bring donations for the animal shelter instead of gifts.’” “Donations can be valuable.” “Dog blankets, Molly. Squeaky toys. Possibly kibble. We’ll be the only burglars in town whose getaway car attracts Labradors.” “We could sell the blankets.” “To whom? Cold dogs?” “I hadn’t got that far.” Bill slid the invitation back. “My sister adopted a dog from that shelter. He sleeps on an old bath mat.” Molly rubbed out the arrow marked EXIT WITH LOOT. “We’ve already rented the magician’s costume,” she said. “So we do the party.” “For free?” “I’ll bring the scarf.” “And I’ll buy a blanket. One blanket. Nobody starts calling us generous.” “What about the goose?” “We ask whether it’s on the guest list.” Bill nodded. “I’m not sharing a dressing room with another bird.”
created_at
"2026-10-05T15:42:20.585Z"
10/5/2026, 3:42:20 PM
model
"openai/gpt-6.1-sol-20260929"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
7572
moderation_latency
(null)
generation_time
28868
tokens_prompt
75
tokens_completion
1012
native_tokens_prompt
66
native_tokens_completion
884
native_tokens_completion_images
0
native_tokens_reasoning
202
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
"completed"
service_tier
"default"
usage
0.008972
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1791214940-5A4RVrjMOrs4ZuuLRhsW"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1791214940-7gmqeuQ2EscYfn2w2281"
upstream_id
"resp_0da8f228a01bfcdd016ac3c55cb5cc87d1a90cc504cdf7dce5"
provider_responses
0
endpoint_id
"dfb9d5c3-63f9-4263-8b78-4fa97dcb3d5e"
id
"resp_0da8f228a01bfcdd016ac3c55cb5cc87d1a90cc504cdf7dce5"
is_byok
false
latency
459
model_permaslug
"openai/gpt-6.1-sol-20260929"
provider_name
"OpenAI"
status
200
total_cost
0.008972
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
99.74%
Matches word count
n/a
neededClean
false
words
504
10.70%
Dialogue to Total Word Ratio
Ratio: 63.31%, Deviation: 6.69%
neededClean
false
wordsTotal
507
wordsDialogue
321
55.2197%