Unverified50% confidenceFactExact time
DeepSeek's official default thinking-effort tier is the Hi tier (short-question thinking cap of 4096 or 8192 tokens); the actual amount of work done by the same-named model under different configurations is not the same.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/20/2026
First Seen
Valid until: 1/18/2027
Sources
Related Entities
Related Claims
UnverifiedDeepSeek Pro's 1.6T total parameters with 49B activated parameters result in inference costs comparable to a roughly 49B dense model, while expanding knowledge capacity by approximately 32 times.74% similarUnverifiedDeepSeek minimal mode paired with Max tends to elicit the model's strongest reasoning capabilities, at the cost of longer response latency and higher token consumption.73% similarUnverifiedModels on DeepSeek's API platform now control the on/off of thinking mode and thinking intensity (low/medium/high) through parameters, rather than releasing two independent models.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1077341API
curl https://kongchang.com/api/v1/knowledge/claims/1077341MCP
get_claim(id=1077341)