Expired50% confidenceFactExact time
Training a model on the scale of GPT-4 requires hundreds of millions of dollars in investment, while the marginal cost of building applications by calling its API has dropped to extremely low levels.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/10/2026
First Seen
Valid until: 10/8/2026(expired)
Sources
AI Engineer大会预演观察：AI工程走向成熟的行业信号
twitterAlexReibman6/25/2026
Related Claims
UnverifiedLarge models (such as GPT-4) can be used to automatically generate synthetic training data from domain documents, reducing annotation costs.78% similarVerified训练一次GPT-4级别的模型需要数万块A100/H100 GPU协同工作数月78% similarUnverifiedAPI fees for calling large language models such as GPT-4 and Claude can accumulate rapidly; it is not uncommon for medium-sized enterprise-level AI applications to consume tens of thousands to hundreds of thousands of dollars in API fees per month during the development and testing phase.78% similarUnverifiedIndustry estimates suggest GPT-4's training used approximately 25,000 NVIDIA A100 GPUs over several months, with compute costs alone exceeding $100 million76% similarUnverifiedGPT-4 completed training in August 2022 with fewer than 2 trillion parameters; the current GPT-5.6 has approximately 4 trillion parameters and Fable approximately 10 trillion parameters, only 5 to 6 times that of GPT-4.76% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/420149API
curl https://kongchang.com/api/v1/knowledge/claims/420149MCP
get_claim(id=420149)