Unverified50% confidenceBenchmarkExact time
In multi-task scenarios, the decision model averages 175.28ms for one question at a time and 191.35ms for two questions at once, a marginal cost of about 16.07ms; attaching 5 classification judgments increases latency by at most 60-80ms.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/19/2026
First Seen
Valid until: 1/17/2027
Sources
Related Entities
Related Claims
UnverifiedThe model offers three reasoning levels: low, high and max; in the author's tests max generated about 16% more tokens than high, with almost the same total time for reasoning tasks.73% similarUnverifiedIn a simulation of 2000 requests, the small model (1 unit cost/request) had an accuracy of 74.5% and a cost of 1.34 per correct answer; the large model (20 unit cost/request) had an accuracy of 95.8% and a cost of 20.88 per correct answer.70% similarUnverifiedThis collection of mathematical results was obtained by posing approximately 4,000 questions to an unreleased internal model, with each result consuming on average about 3 hours of the model's thinking compute on ChatGPT Pro.70% similarUnverifiedBy selecting the appropriate model for different tasks, the solution rate improved by more than 23 percentage points, and the per-task cost was reduced to less than one-third of the original.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1069968API
curl https://kongchang.com/api/v1/knowledge/claims/1069968MCP
get_claim(id=1069968)