feat: add 2026 Q2 flagship models (#76)

Add latest flagships released in April 2026 that registry missed:
- deepseek: V4-Pro, V4-Flash (2026-04-24, 1M context)
- qwen: qwen3.6-max-preview (2026-04-20, 256K context)
- moonshot: kimi-k2.6 (2026-04-20, 256K context)
- zhipu: glm-5.1 (2026-04-08), glm-4.7-flash (free tier)

Update default aliases: deepseek -> v4-pro, kimi -> k2.6, glm -> 5.1.

providers/deepseek.toml

@@ -1,5 +1,5 @@
 # DeepSeek — https://deepseek.com
-# Models: 2
+# Models: 4
 
 [provider]
 id = "deepseek"
@@ -8,6 +8,34 @@ api_key_env = "DEEPSEEK_API_KEY"
 base_url = "https://api.deepseek.com/v1"
 key_required = true
 
+[[models]]
+id = "deepseek-v4-pro"
+display_name = "DeepSeek V4 Pro"
+tier = "frontier"
+context_window = 1000000
+max_output_tokens = 384000
+input_cost_per_m = 0.435
+output_cost_per_m = 0.87
+supports_tools = true
+supports_vision = false
+supports_streaming = true
+supports_thinking = true
+aliases = ["deepseek-v4", "deepseek-pro"]
+
+[[models]]
+id = "deepseek-v4-flash"
+display_name = "DeepSeek V4 Flash"
+tier = "smart"
+context_window = 1000000
+max_output_tokens = 384000
+input_cost_per_m = 0.14
+output_cost_per_m = 0.28
+supports_tools = true
+supports_vision = false
+supports_streaming = true
+supports_thinking = true
+aliases = ["deepseek-flash"]
+
 [[models]]
 id = "deepseek-chat"
 display_name = "DeepSeek V3"