New deepseek limits and prices

[stefanmohl]

Issue

Since Deepseek 3.2 was released:

#filename="~/.aider.model.metadata.json"

{
    "deepseek/deepseek-chat": {
        "max_tokens": 128000,
        "max_input_tokens": 128000,
        "max_output_tokens": 8000,
        "input_cost_per_token": 0.28e-6,
        "input_cost_per_token_cache_hit": 0.028e-6,
        "output_cost_per_token": 0.42e-6,
    },
    "deepseek/deepseek-reasoner": {
        "max_tokens": 128000,
        "max_input_tokens": 128000,
        "max_output_tokens": 64000,
        "input_cost_per_token": 0.28e-6,
        "input_cost_per_token_cache_hit": 0.028e-6,
        "output_cost_per_token": 0.42e-6,
    }
}

Version and model info

No response

[dwash96]

Thank you for this, I was going to update with the base repo today, but glad to see another extremely common deepseek W

[stefanmohl]

Oh! I am really sorry but I just realised that I switched the input cost for cache hit and cache miss with each other! I have edited my post to be correct now.