NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-10-05-01M46BKST62RRT18RT639B81QB
Run Details
Model:
GPT-6.1 Sol (Reasoning, Medium)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
1m 6s
Cost:
$0.029752
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
99.98%
Dialogue to Total Word Ratio
(details)
99.9910%
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 the birthday invitation across their table, flattening a glittery unicorn with her elbow. “We’re stealing presents from a children’s party. We need dignity.” “Then why did you buy matching raccoon masks?” “Branding, Bill. People trust consistency, even when consistency involves criminals dressed like deeply disappointed suburban woodland accountants.” Bill inspected the invitation through his reading glasses, which made him resemble a fraudulent dentist. “It says fancy dress.” “Exactly. You’re the magician.” “I can’t do magic.” “You made our rent disappear.” “That was cryptocurrency.” “Same thing, but magicians usually get invited back for cake afterward by surprisingly forgiving people.” Molly turned over a notebook, revealing diagrams of balloons and one extremely judgmental stick figure. “Who’s that?” “The birthday child.” “Why does she have a sword?” “It’s a balloon.” “Why does the balloon have blood on it?” “That’s jam. Stop making everything sinister. We’re professionals, not dramatists with childhood trauma.” Bill opened their equipment bag and extracted rubber gloves, a shopping sack, and emergency antacids. “We take everything?” “Every present.” “Even educational toys?” “Especially those. Rich parents pay fortunes for wooden objects that teach children disappointment.” “What about the cards?” “Leave them. We’re thieves, Bill, not actual completely irredeemable monsters.” A party horn rolled across the table. Bill watched it with mistrust reserved for snakes. “How do we avoid the parents?” “Confidence. Carry a clipboard.” “What’s on it?” “A list of allergies.” “Whose allergies?” “Ours. You’re allergic to confrontation, and I develop hives when asked to fund communal retirement gifts.” Molly produced a photograph of the venue. Bill leaned closer, then stopped chewing his pencil. “That’s my sister’s house.” “You said your sister lived somewhere terrible.” “Yes. Suburbia.” “You know the birthday child?” “She’s my niece.” “Then we’ve got inside knowledge.” “She’s six. She knows all my deepest personal failures.” The unicorn on the invitation seemed suddenly smug. Molly covered its face with a coaster. “Fine. We cancel.” “We can’t. I already promised to come.” “As a magician?” “As a responsible adult.” “That’s a much harder trick.” “She specifically requested you, too.” “Why?” “Apparently, you look like her cartoon villain.” Bill rummaged through the bag again, retrieving two unopened toys purchased for their cover story. “We bought decoy presents.” “Technically, investments.” “Could we simply give her these?” “And leave empty-handed?” “They’ll have cake.” “Molly, that’s not stealing.” “It is if we take more than our share and conceal slices elsewhere.” Molly crossed out the heist plan and wrote CAKE beneath it in determined capital letters. “We still need an escape strategy.” “Before the cleanup.” “And if your sister catches us?” “Say you have another booking.” “Doing what?” “Magic.” “Bill, I cannot stress enough how little magic either of us possesses.” Bill pocketed the invitation while Molly reluctantly returned their raccoon masks to the equipment bag. “What if she asks us to help wash dishes?” “We vanish.” “Finally, a trick we can perform.” “Together?” “Obviously. Partners.” “And the presents?” “Let the kid keep them. We’ll rob somebody taller, with less glitter.”
created_at
"2026-10-05T15:41:12.401Z"
10/5/2026, 3:41:12 PM
model
"openai/gpt-6.1-sol-20260929"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
8540
moderation_latency
(null)
generation_time
65886
tokens_prompt
75
tokens_completion
1288
native_tokens_prompt
66
native_tokens_completion
2962
native_tokens_completion_images
0
native_tokens_reasoning
2264
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
0.029752
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1791214872-wPmEGhRLseo6Luo0Pa7b"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1791214872-eBjaiwn1HMHkp40zCU9b"
upstream_id
"resp_073aafe5895f6940016ac3c51889fc87d196f34af48b43d42a"
provider_responses
0
endpoint_id
"dfb9d5c3-63f9-4263-8b78-4fa97dcb3d5e"
id
"resp_073aafe5895f6940016ac3c51889fc87d196f34af48b43d42a"
is_byok
false
latency
335
model_permaslug
"openai/gpt-6.1-sol-20260929"
provider_name
"OpenAI"
status
200
total_cost
0.029752
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
100.00%
Matches word count
n/a
neededClean
false
words
500
99.98%
Dialogue to Total Word Ratio
Ratio: 70.06%, Deviation: 0.06%
neededClean
false
wordsTotal
501
wordsDialogue
351
99.9910%