BREAKING
DeepSeek Cuts Cached Input 10x
0$/M
cached hit
0$/M
cache miss
0$/M
output
0%
V4-Pro discount
0T
total params
0B
active params
Why It Matters for Agents
1Repeated context
2Prefix cache hit
3Tiny input cost
4Prompt change misses
DeepSeek vs Grok 4.5
DeepSeek V4-Pro~10x cheaper
Cached ~$0.05/M effective
1M-token context
Grok 4.5July 2026
Input $2, output $6 per M
Cached ~25% of standard
Cache Efficiency Fuels Price War
AI NEWS BLITZ
DeepSeek has slashed cached input token prices tenfold across all its API models.