Repository navigation
Model context limits not respected #11286
Description
Description
opencode.jsonc has:
{
"$schema": "https://opencode.ai/config.json",
"provider": {
"llama.cpp": {
"npm": "@ai-sdk/openai-compatible",
"name": "Llama.cpp (local)",
"options": {
"baseURL": "http://127.0.0.1:8080/v1",
"timeout": false
},
"models": {
"minimax-m2.1-reap-30": {
"name": "Minimax M2.1 REAP-30",
"limit": {
"context": 65536,
"output": 16384
}
}
}
}
}
}
Somewhere along the way, the model tries to read a file, but the file is rather large:
→ Read file.py
┃
┃ request (68879 tokens) exceeds the available context size (65536 tokens), try increasing it
┃
That error is from llama.cpp:
slot update_slots: id 3 | task 0 | new prompt, n_ctx_slot = 65536, n_keep = 0, task.n_tokens = 68885
srv send_error: task id = 0, error: request (68885 tokens) exceeds the available context size (65536 tokens), try increasing it
Which means this is an opencode bug. It should report back that the file can't be read due to context limits.
Plugins
No response
OpenCode version
1.1.40
Steps to reproduce
No response
Screenshot and/or share link
No response
Operating System
Ubuntu 24.04
Terminal
iTerm2
Reacted by Damian S, Kevin Cox, peltax, Mateo Bohorquez G. and Vladimir
Reacted by Mateo Bohorquez G.
Activity
github-actions commentedon Jan 30, 2026
This issue might be a duplicate of existing issues. Please check:
Feel free to ignore if none of these address your specific case.
github-actions commentedon Apr 1, 2026
To stay organized issues are automatically closed after 90 days of no activity. If the issue is still relevant please open a new one.
Milor123 commentedon Jun 17, 2026
@gcp what is the solucion?
gcp commentedon Jun 17, 2026
I have no idea, probably "don't use opencode".
Milor123 commentedon Jun 17, 2026
Ohh no, I found the solution for llama.cpp
The parameter parallel divides the context into the normal, I've changed from
-np 4to-np 1and it was solved for me using llama.cpp✅