Give qwen3-coder a 128k context

At 32k, opencode auto-compacted the history four times in four minutes
of reading files; after each summary the model lost the task, and after
the last it started writing tool calls as JSON text. 128k fits in ~31 GB.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

README.md

@@ -21,7 +21,7 @@ Needs `brew install ollama opencode`.
 
 | Model | From | Used for |
 |-------|------|----------|
-| `qwen3-coder:30b-32k` | `qwen3-coder:30b` | opencode's default |
+| `qwen3-coder:30b-128k` | `qwen3-coder:30b` | opencode's default (128k context) |
 | `qwen3:32b-32k` | `qwen3:32b` | opencode, as an option |
 | `inspector` | `qwen3:32b` | 16k-context variant |
 

ollama/models/qwen3-coder-30b-128k.Modelfile

@@ -0,0 +1,7 @@
+# name: qwen3-coder:30b-128k
+# opencode's default: trained for agentic tool calling, no thinking phase,
+# mixture-of-experts (~3B active params) so it's fast. 128k context: at 32k,
+# opencode compacted the history every minute or so of file reading, and the
+# model lost track of the task after each summary. ~30 GB loaded.
+FROM qwen3-coder:30b
+PARAMETER num_ctx 131072

ollama/models/qwen3-coder-30b-32k.Modelfile

@@ -1,5 +0,0 @@
-# name: qwen3-coder:30b-32k
-# opencode's default: trained for agentic tool calling, no thinking phase,
-# mixture-of-experts (~3B active params) so it's fast.
-FROM qwen3-coder:30b
-PARAMETER num_ctx 32768

opencode/opencode.json

@@ -8,19 +8,25 @@
         "baseURL": "http://localhost:11434/v1"
       },
       "models": {
-        "qwen3-coder:30b-32k": {
-          "name": "Qwen3 Coder 30B (32k context)",
+        "qwen3-coder:30b-128k": {
+          "name": "Qwen3 Coder 30B (128k context)",
           "tools": true,
-          "limit": { "context": 32768, "output": 8192 }
+          "limit": {
+            "context": 131072,
+            "output": 8192
+          }
         },
         "qwen3:32b-32k": {
           "name": "Qwen3 32B (32k context)",
           "tools": true,
-          "limit": { "context": 32768, "output": 8192 }
+          "limit": {
+            "context": 32768,
+            "output": 8192
+          }
         }
       }
     }
   },
-  "model": "ollama/qwen3-coder:30b-32k",
-  "small_model": "ollama/qwen3-coder:30b-32k"
+  "model": "ollama/qwen3-coder:30b-128k",
+  "small_model": "ollama/qwen3-coder:30b-128k"
 }