Unverified50% confidenceFactExact time
Reasoning models are extremely costly to train, typically requiring a combination of large-scale reinforcement learning and carefully designed reward models to acquire the capability.
1
Sources
50%
Confidence
Long-term
Relevance
9/30/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedThe energy consumption during the training phase of large models is approximately hundreds of times that of the inference phase.80% similarVerifiedTraining frontier large models requires enormous computational investment, and there are diminishing marginal returns as model scale increases.79% similarUnverified训练2.8万亿参数模型代价高昂，便宜和开源并不等于不需要算力，算力需求从闭源巨头向分散的企业定制场景转移78% similarUnverifiedThe cost pressures on frontier large model companies primarily stem from compute chip procurement for model training and compensation for top research talent78% similarUnverifiedReinforcement learning-based methods require large amounts of data and repeated training, resulting in high training costs and limited adaptability to environmental changes.76% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/962777API
curl https://kongchang.com/api/v1/knowledge/claims/962777MCP
get_claim(id=962777)