Opencode - Historique des versions

Mathieu le 18 mars 2026 à 22:50

2026-03-18T22:50:37Z

@@ Ligne 32 : / Ligne 32 : @@
 NAME              ID              SIZE      PROCESSOR    CONTEXT    UNTIL
 qwen3.5:2b-32k    094e78c5fe51    5.1 GB    100% GPU     32768      4 minutes from now
 </syntaxhighlight>

Mathieu le 18 mars 2026 à 22:07

2026-03-18T22:07:24Z

← Version précédente		Version du 19 mars 2026 à 00:07
Ligne 2 :		Ligne 2 :
	https://ollama.com/search list model size on disk, not on VRAM.		https://ollama.com/search list model size on disk, not on VRAM.

	This site list some vram usage : https://localllm.in/blog/ollama-vram-requirements-for-local-llms		This site list some vram usage : https://localllm.in/blog/ollama-vram-requirements-for-local-llms or https://github.com/ollama/ollama/issues/6852#issuecomment-2440229918

	https://www.canirun.ai/model/llama3.1-8b may help but Ollama may not give access to all quatization		https://www.canirun.ai/model/llama3.1-8b may help but Ollama may not give access to all quatization

Mathieu : /* Find a suitable model that fit in RAM */

2026-03-18T19:36:10Z

Find a suitable model that fit in RAM

← Version précédente		Version du 18 mars 2026 à 21:36
Ligne 1 :		Ligne 1 :
	=== Find a suitable model that fit in RAM ===		=== Find a suitable model that fit in RAM ===
	https://ollama.com/search list model size on disk, not on VRAM.		https://ollama.com/search list model size on disk, not on VRAM.

			This site list some vram usage : https://localllm.in/blog/ollama-vram-requirements-for-local-llms

	https://www.canirun.ai/model/llama3.1-8b may help but Ollama may not give access to all quatization		https://www.canirun.ai/model/llama3.1-8b may help but Ollama may not give access to all quatization

Mathieu : /* Increase model's context size */

2026-03-18T10:22:07Z

Increase model's context size

@@ Ligne 22 : / Ligne 22 : @@
 Created new model 'qwen3.5:2b-32k'
 </syntaxhighlight>
 ==== Check if it fits in VRAM ====
 Run the model:<syntaxhighlight lang="bash">

Mathieu : /* Increase model's context size */

2026-03-18T10:17:06Z

Increase model's context size

@@ Ligne 4 : / Ligne 4 : @@
 https://www.canirun.ai/model/llama3.1-8b may help but Ollama may not give access to all quatization
-Also find a model that can think and run tools ( this may help https://www.canirun.ai/ )
+Also find a model that can run tools (and think?) ( this may help https://www.canirun.ai/ )
@@ Ligne 22 : / Ligne 22 : @@
 Created new model 'qwen3.5:2b-32k'
 </syntaxhighlight>

Mathieu le 18 mars 2026 à 10:00

2026-03-18T10:00:29Z

@@ Ligne 1 : / Ligne 1 : @@
 === Find a suitable model that fit in RAM ===
 https://ollama.com/search list model size on disk, not on VRAM.
 Also find a model that can think and run tools ( this may help https://www.canirun.ai/ )
@@ Ligne 52 : / Ligne 54 : @@
 </syntaxhighlight>

Mathieu : Page créée avec « === Find a suitable model that fit in RAM === https://ollama.com/search list model size on disk, not on VRAM. Also find a model that can think and run tools ( this may help https://www.canirun.ai/ ) Model take VRAM size, but context size also. As stated here, we may have to increase context size to 16k-32k. https://opencode.ai/docs/fr/providers/#ollama ==== Increase model's context size ==== ollama run qwen3.5:2b >>> /set pa... »

2026-03-18T09:54:05Z

Page créée avec « === Find a suitable model that fit in RAM === https://ollama.com/search list model size on disk, not on VRAM. Also find a model that can think and run tools ( this may help https://www.canirun.ai/ ) Model take VRAM size, but context size also. As stated here, we may have to increase context size to 16k-32k. https://opencode.ai/docs/fr/providers/#ollama ==== Increase model's context size ==== <syntaxhighlight lang="bash"> ollama run qwen3.5:2b >>> /set pa... »

Nouvelle page

=== Find a suitable model that fit in RAM ===
https://ollama.com/search list model size on disk, not on VRAM.

Also find a model that can think and run tools ( this may help https://www.canirun.ai/ )

Model take VRAM size, but context size also. As stated here, we may have to increase context size to 16k-32k.

https://opencode.ai/docs/fr/providers/#ollama

==== Increase model's context size ====
<syntaxhighlight lang="bash">
ollama run qwen3.5:2b

>>> /set parameter num_ctx 32768

Set parameter 'num_ctx' to '32768'

>>> /save qwen3.5:2b-32k

Created new model 'qwen3.5:2b-32k'
</syntaxhighlight>

==== Check if it fits in VRAM ====
Run the model:<syntaxhighlight lang="bash">
ollama run qwen3.5:2b-32k
</syntaxhighlight>Then check CPU vs GPU usage. It should be 100% GPU to stay fast<syntaxhighlight lang="bash">
ollama ps
NAME ID SIZE PROCESSOR CONTEXT UNTIL
qwen3.5:2b-32k 094e78c5fe51 5.1 GB 100% GPU 32768 4 minutes from now
</syntaxhighlight>

=== Configure Opencode ===
in ~/.config/opencode/config.json<syntaxhighlight lang="json">
{
"$schema": "https://opencode.ai/config.json",
"provider": {
"ollama": {
"name": "Ollama (spacemarine)",
"npm": "@ai-sdk/openai-compatible",
"options": {
"baseURL": "http://192.168.0.24:11434/v1"
},
"models": {
"qwen3.5:2b-32k": {
"tools": true
}
}
}
}
}

</syntaxhighlight>

Opencode - Historique des versions

Mathieu le 18 mars 2026 à 22:50

Mathieu le 18 mars 2026 à 22:07

Mathieu : /* Find a suitable model that fit in RAM */

Mathieu : /* Increase model's context size */

Mathieu : /* Increase model's context size */

Mathieu le 18 mars 2026 à 10:00

Mathieu : /* Find a suitable model that fit in RAM */

Mathieu : /* Increase model's context size */

Mathieu : /* Increase model's context size */