Unverified50% confidenceFactExact time
This project's fact extraction pipeline runs inference three times per session rather than once, making the cost three times that of a normal pipeline.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/14/2026
First Seen
Valid until: 1/12/2027
Sources
Related Entities
Related Claims
UnverifiedEvaluating the per-task cost of large models requires considering both the token unit price and the number of tokens consumed to complete the task; the output token consumption of chain-of-thought reasoning models may increase by several times or even dozens of times.66% 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.66% 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.65% similarUnverifiedA third-party test classifying 1,000 emails found the judgment model took 15.6 seconds at a cost of 1.77, while the large model took 48.9 seconds at a cost of 2.07—approximately 15% lower cost but roughly three times faster.64% similarUnverifiedReasoning models may consume several times or even more than ten times the output tokens, directly driving up API call costs.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1023340API
curl https://kongchang.com/api/v1/knowledge/claims/1023340MCP
get_claim(id=1023340)