DeepSeek

14 models compared CN

DeepSeek currently has 14 models listed on LLM Meter. Input prices range from $0.10 to $1.42 and output prices from $0.21 to $2.83 per 1M tokens; the largest context window is 1,048,576 tokens. 2 of them are priced from DeepSeek’s own API; the rest show a median third-party host price (≈).

Cheapest model
$0.10/M input
Largest context
1,048,576 tokens
Model Input ($/1M) Output ($/1M) Context Features Free
DeepSeek V4 Flash 0731 ≈$0.10 ≈$0.42 1.05M function-callingjson-modethinking -
DeepSeek V4 Flash 0423 ≈$0.10 ≈$0.21 1.05M function-callingjson-modethinking -
DeepSeek V4.1 Flash 1017 $0.15 $0.60 1.05M visionfunction-callingjson-mode +1 -
DeepSeek V4.1 Flash ≈$0.21 ≈$0.63 1.05M visionfunction-callingjson-mode +1 -
DeepSeek V4 Flash Vision Exp ≈$0.22 ≈$0.65 1.05M visionfunction-callingjson-mode +1 -
DeepSeek V3 0324 ≈$0.25 ≈$1.00 164K function-callingjson-mode -
DeepSeek V3 ≈$0.26 ≈$1.03 164K function-callingjson-mode -
DeepSeek V3.2 Exp ≈$0.27 ≈$0.41 164K function-callingjson-modethinking -
DeepSeek V3.1 Terminus ≈$0.27 ≈$1.00 164K function-callingjson-modethinking -
DeepSeek V3.1 ≈$0.27 ≈$1.00 164K function-callingjson-modethinking -
DeepSeek V3.2 ≈$0.28 ≈$0.42 164K function-callingjson-modethinking -
R1 0528 ≈$0.50 ≈$2.18 164K function-callingjson-modethinking -
DeepSeek V4 Pro 0813 $0.66 $1.98 1.05M function-callingjson-modethinking -
DeepSeek V4 Pro 0423 ≈$1.42 ≈$2.83 1.05M function-callingjson-modethinking -

No longer listed (6)

R1, R1 Distill Llama 70B, DeepSeek V4 Pro, DeepSeek V4 Flash, R1 Distill Qwen 32B, DeepSeek V3.2 Speciale