ModelCap
Current AI model comparison
Exact identities · public evidence
DeepSeek
DeepSeek V4.1 Flash 1017
Position
#70
Index
61.5
Measured
publisher-corpus-prior over 210 held-out anchors (18%); measured predecessor deepseek/deepseek-v4-flash less the succession penalty (69%); 1 specialist-board observation at 2.1% support (13%); shrunk 0.4 toward the measured corpus
Comparison side 1
DeepSeek
DeepSeek V4 Pro 0813
Position
#29
Index
74.9
Measured
3 public benchmark observations across 3 boards
Comparison side 2
Output / 1M
$0.60
vs $1.98
Context
1M
vs 1M
Weights
Open weights
vs Open weights