NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M37HTQY5A44FSYK7SW08GGR6
Run Details
Model:
GPT-6 Sol (Reasoning, Medium)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
20.5s
Cost:
$0.008392
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
60.28%
Matches word count
(details)
10.59%
Dialogue to Total Word Ratio
(details)
35.4329%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 30% 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
From the roof of the library, Molly studied the birthday party through a pair of opera glasses. Below, twenty children chased balloons around a courtyard. A mountain of wrapped presents occupied the picnic table. At its summit sat a box the size of a washing machine, tied with a bow that trembled in the breeze. “We go in as entertainers,” Molly said. “Absolutely not,” said Bill. “Last time I was an entertainer, someone made me entertain.” “That’s generally the risk.” “A girl asked me to pull a rabbit from my hat. I didn’t have a rabbit. I didn’t have a hat.” “You had my handbag.” “Yes. And now she has your car keys.” Molly lowered the glasses. Bill was wearing a black turtleneck, black trousers, and a paper crown from a burger restaurant. He had insisted the crown made him look approachable. It made him look like the king of a very small, troubled country. “Fine. We go in as parents,” she said. “Whose parents?” “Anyone’s. Children don’t pay attention.” “Parents do.” “They’re busy cutting cake and wondering why they had children.” Below, a man emerged from the library carrying a tray of juice cartons. He stopped to retrieve a balloon from a tree, lost his grip on the tray, and stood helplessly while a dozen children descended upon the wreckage. The presents, momentarily, had no guard. Molly pointed. Bill leaned so far over the parapet that his paper crown sailed down into the courtyard. It landed squarely atop the largest present. “There goes our approachable disguise,” he said. “Maybe it’s an improvement.” “What’s the big box?” “Bicycle, possibly. Or a miniature horse.” “That bow isn’t big enough for a horse.” “You know less about horses than you think.” They watched a little boy approach the mountain. Rather than pick up a gift, he checked a clipboard and moved three parcels into a neat row. Another child brought him a fourth. Molly adjusted the glasses. On every label, in careful block capitals, was the same destination: the children’s hospital. Bill noticed her expression. “What?” “They’re collecting presents. For other children.” “Oh.” He considered this. “That makes us worse than usual.” “Considerably.” “We could steal them and deliver them ourselves.” “That would be the same delivery with unnecessary crime.” Bill looked disappointed. He had spent all morning perfecting a dramatic exit, which involved a cape, a skateboard, and what he called “the element of downhill.” Molly was about to suggest they leave when the clipboard boy spotted the crown. He lifted the lid of the enormous box. Out sprang a cardboard dragon operated by six giggling children, its jaws opening wide enough to swallow the crown. Bill gripped Molly’s sleeve. “I have a new plan.” “If it involves the dragon, no.” “It involves the cake.” “You’re paying for it.” “Then it’s not a heist,” Bill said. “No,” Molly replied, heading for the stairs. “It’s a party.”
created_at
"2026-09-23T16:33:23.918Z"
9/23/2026, 4:33:23 PM
model
"openai/gpt-6-sol-20260922"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
7481
moderation_latency
(null)
generation_time
20511
tokens_prompt
75
tokens_completion
806
native_tokens_prompt
66
native_tokens_completion
826
native_tokens_completion_images
0
native_tokens_reasoning
193
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
0.008392
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.20.0; linux; x64))"
http_referer
(null)
request_id
"req-1790181203-bQboN9WGMbAGMDWSLJTK"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790181203-vlalwDFu465QKmeXlyOI"
upstream_id
"resp_0e7e88e077eb8f42016ab3ff54630887d182e543840b3d9576"
provider_responses
0
endpoint_id
"1cfc7d9d-4404-4b8e-9ee4-58ae45c9dcd4"
id
"resp_0e7e88e077eb8f42016ab3ff54630887d182e543840b3d9576"
is_byok
false
latency
1616
model_permaslug
"openai/gpt-6-sol-20260922"
provider_name
"OpenAI"
status
200
total_cost
0.008392
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
60.28%
Matches word count
n/a
neededClean
false
words
485
10.59%
Dialogue to Total Word Ratio
Ratio: 36.70%, Deviation: 6.70%
neededClean
false
wordsTotal
485
wordsDialogue
178
35.4329%