Revert "fix max_tokens for reasoning models (#1819)" (#1839)

* Revert "fix max_tokens for reasoning models (#1819)"

This reverts commit 1e2e233ff5.

* Revert "fix: revert max_completion_tokens to max_tokens (#1741)"

This reverts commit cd13eeb7d9.

* fix: nvim_version

README.md

@@ -69,7 +69,7 @@ For building binary if you wish to build from source, then `cargo` is required.
       model = "gpt-4o", -- your desired model (or use gpt-4o, etc.)
       timeout = 30000, -- Timeout in milliseconds, increase this for reasoning models
       temperature = 0,
-      max_tokens = 8192, -- Increase this to include reasoning tokens (for reasoning models)
+      max_completion_tokens = 8192, -- Increase this to include reasoning tokens (for reasoning models)
       --reasoning_effort = "medium", -- low|medium|high, only used for reasoning models
     },
   },

cursor-planning-mode.md

@@ -35,7 +35,7 @@ Then enable it in avante.nvim:
             api_key_name = 'GROQ_API_KEY',
             endpoint = 'https://api.groq.com/openai/v1/',
             model = 'llama-3.3-70b-versatile',
-            max_tokens = 32768, -- remember to increase this value, otherwise it will stop generating halfway
+            max_completion_tokens = 32768, -- remember to increase this value, otherwise it will stop generating halfway
         },
     },
     --- ... existing configurations

lua/avante/config.lua

@@ -219,7 +219,7 @@ M._defaults = {
     model = "gpt-4o",
     timeout = 30000, -- Timeout in milliseconds, increase this for reasoning models
     temperature = 0,
-    max_tokens = 16384, -- Increase this to include reasoning tokens (for reasoning models)
+    max_completion_tokens = 16384, -- Increase this to include reasoning tokens (for reasoning models)
     reasoning_effort = "medium", -- low|medium|high, only used for reasoning models
   },
   ---@type AvanteSupportedProvider
@@ -239,7 +239,7 @@ M._defaults = {
     api_version = "2024-12-01-preview",
     timeout = 30000, -- Timeout in milliseconds, increase this for reasoning models
     temperature = 0,
-    max_tokens = 20480, -- Increase this to include reasoning tokens (for reasoning models)
+    max_completion_tokens = 20480, -- Increase this to include reasoning tokens (for reasoning models)
     reasoning_effort = "medium", -- low|medium|high, only used for reasoning models
   },
   ---@type AvanteSupportedProvider

lua/avante/providers/azure.lua

@@ -2,7 +2,7 @@
 ---@field deployment string
 ---@field api_version string
 ---@field temperature number
----@field max_tokens number
+---@field max_completion_tokens number
 ---@field reasoning_effort? string
 
 local Utils = require("avante.utils")

lua/avante/providers/openai.lua

@@ -67,13 +67,8 @@ function M.set_allowed_params(provider_conf, request_body)
   else
     request_body.reasoning_effort = nil
   end
-
-  if M.is_reasoning_model(provider_conf.model) then
-    if request_body.max_tokens then
-      request_body.max_completion_tokens = request_body.max_tokens
-      request_body.max_tokens = nil
-    end
-  end
+  -- If max_tokens is set in config, unset max_completion_tokens
+  if request_body.max_tokens then request_body.max_completion_tokens = nil end
 end
 
 function M:parse_messages(opts)

lua/avante/types.lua

@@ -217,6 +217,7 @@ vim.g.avante_login = vim.g.avante_login
 ---@field __inherited_from? string
 ---@field temperature? number
 ---@field max_tokens? number
+---@field max_completion_tokens? number
 ---@field reasoning_effort? string
 ---@field display_name? string
 ---