Can OpenCode use fully local open-source models?

Yes. OpenCode connects to any OpenAI-compatible endpoint — Ollama (localhost:11434/v1), llama-server (localhost:8080/v1), LM Studio, or vLLM. Add a custom provider in ~/.config/opencode/opencode.jsonc with baseURL and apiKey set to local, then set "model" to your provider/model id. No code leaves your machine if the inference server runs locally.

What is the easiest local stack for OpenCode in 2026?

For beginners: Ollama pull + ollama serve, then OpenCode /connect or a two-line opencode.jsonc pointing at http://127.0.0.1:11434/v1. For maximum control and speed on Apple Silicon: llama.cpp with MTP (see Qwen 3.6 27B guide). For GUI model browsing: LM Studio local server mode.

Which open-source model should I run locally for coding with OpenCode?

Dense Qwen 3.6 27B is the current HN sweet spot for instruction-following and agent tasks on 48GB+ unified memory. GLM-5.2 via Unsloth needs heavier hardware but closer to frontier. Gemma 4 12B/31B if you need multimodal local. Match quant to VRAM — Q8 on 48GB Mac, Q4 on 32GB. See Mac vs GPU guide for tiers.

Can I mix local models and cloud APIs in OpenCode?

Yes. Define multiple providers in opencode.jsonc — local llama.cpp for private repos, Z.AI GLM Coding Plan or OpenRouter for frontier bursts. Switch with /models or edit the default "model" field. Tier workloads like enterprise Fable-alternative guides recommend: sensitive code local, peak reasoning on API.

What context window does OpenCode need for local models?

Agent coding loops benefit from 32k–64k minimum; 128k+ helps long refactors. Qwen 3.6 supports 256k native but 64k is a practical RAM default. Codex OSS mode documents 64k as a floor for agent loops — treat that as a good minimum for OpenCode local sessions too.

Run Open Source Models Locally in OpenCode (2026) | explainx.ai Blog

Q: How do I configure OpenCode for a local llama.cpp server?

Start llama-server on port 8080 with your GGUF model, then add to opencode.jsonc a provider block with npm @ai-sdk/openai-compatible, options.baseURL http://127.0.0.1:8080/v1, apiKey local, and models map. Set "model": "llama/your-model-id". Full copy-paste in this guide and the Qwen 3.6 27B post.

July 1, 2026: With Fable 5 still offline, GLM-5.2 and Qwen 3.6 27B proving local coding is practical, the missing piece for many developers is not which model — it is how to wire inference into a harness.

OpenCode is the open-source agent loop that accepts any OpenAI-compatible API. Run weights on your CPU/GPU, expose http://127.0.0.1:…/v1, point ~/.config/opencode/opencode.jsonc at it, and you have a local coding agent — same tools, LSP, and /init AGENTS.md flow as cloud setups, without sending repo context to a third party.

This is explainx.ai's end-to-end stack guide: pick a model → start a server → configure OpenCode → tier local vs cloud.

TL;DR — the full local + OpenCode path

Step	What to do	Deep dive
1. Hardware	32GB+ for Q4 7B–27B; 48GB+ for Q8 27B	Mac vs GPU guide
2. Model	Qwen 3.6 27B coding default; GLM-5.2 if you have GPU headroom	Qwen local guide · GLM Unsloth
3. Runtime	llama.cpp (control) or Ollama (ease)	What is llama.cpp? · Personal AI system
4. Server	OpenAI-compatible `/v1` on localhost	Sections below
5. OpenCode	`opencode.jsonc` provider + `opencode` in repo	OpenCode harness guide
6. Verify	`/init`, small coding task, check tok/s	Quantization guide

Architecture — three layers

text

┌─────────────────────────────────────┐
│  OpenCode (agent harness)           │
│  tools · LSP · AGENTS.md · sessions │
└──────────────┬──────────────────────┘
               │ OpenAI-compatible HTTP
┌──────────────▼──────────────────────┐
│  Inference server                   │
│  llama.cpp · Ollama · LM Studio · vLLM│
└──────────────┬──────────────────────┘
               │ GGUF / safetensors
┌──────────────▼──────────────────────┐
│  Open-weight model on your disk/GPU │
│  Qwen · GLM · Gemma · DeepSeek · …  │
└─────────────────────────────────────┘

OpenCode never loads weights itself — it only talks HTTP. That separation is why one config file can swap Ollama today and llama.cpp tomorrow.

Step 1 — Install OpenCode

bash

curl -fsSL https://opencode.ai/install | bash
# or: npm install -g opencode-ai
# or: brew install anomalyco/tap/opencode

bash

cd your-project
opencode

First-run inside TUI:

/connect — for cloud providers (Z.AI, OpenRouter, Copilot OAuth)
/init — generate AGENTS.md project memory
/models — pick active model after providers exist

Slash reference: OpenCode commands. Harness concepts: What is an agent harness?.

Step 2 — Pick an inference runtime

Runtime	Best for	OpenAI API default	OpenCode fit
llama.cpp	Max control, MTP, Apple Silicon	`http://127.0.0.1:8080/v1`	Recommended for serious local coding
Ollama	One-command pull/serve	`http://127.0.0.1:11434/v1`	Fastest onboarding
LM Studio	GUI + local server toggle	`http://127.0.0.1:1234/v1`	Non-terminal users
vLLM	Multi-user / production GPU box	custom port `/v1`	Team LAN server

Codex users: same local servers work in Codex OSS mode — different harness, identical inference layer.

Step 3 — Start the local server

Option A — llama.cpp (Qwen 3.6 27B example)

From the Qwen 3.6 local benchmark post:

bash

llama-server -hf unsloth/Qwen3.6-27B-MTP-GGUF:Q8_0 \
    --spec-type draft-mtp -ngl 999 -fa on -c 65536 --port 8080

Smoke test: open http://127.0.0.1:8080 or curl http://127.0.0.1:8080/v1/models.

Option B — Ollama

bash

ollama pull qwen3.6:27b        # verify tag on ollama.com/library
ollama serve                   # usually auto-starts on 11434

Other coding pulls teams use: glm-5.2, deepseek-coder-v2, gemma4:12b — check VRAM before pulling 70B-class tags.

Option C — LM Studio

Download model in GUI
Local Server tab → Start Server (default 1234)
Enable OpenAI compatible API

Option D — vLLM (Linux GPU server)

bash

python -m vllm.entrypoints.openai.api_server \
  --model Qwen/Qwen3.6-27B-Instruct \
  --port 8000

Point OpenCode at http://127.0.0.1:8000/v1 or your LAN IP for a shared team box.

Step 4 — Configure OpenCode (`opencode.jsonc`)

Config path: ~/.config/opencode/opencode.jsonc (user-wide) or project-local per OpenCode docs.

llama.cpp provider block

jsonc

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "llama": {
      "name": "llama.cpp (local)",
      "npm": "@ai-sdk/openai-compatible",
      "options": {
        "baseURL": "http://127.0.0.1:8080/v1",
        "apiKey": "local"
      },
      "models": {
        "qwen3.6-27b": { "name": "Qwen3.6-27B Q8 +MTP" }
      }
    }
  },
  "model": "llama/qwen3.6-27b"
}

Ollama provider block

jsonc

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "ollama": {
      "name": "Ollama (local)",
      "npm": "@ai-sdk/openai-compatible",
      "options": {
        "baseURL": "http://127.0.0.1:11434/v1",
        "apiKey": "ollama"
      },
      "models": {
        "qwen3.6-27b": { "name": "qwen3.6:27b" }
      }
    }
  },
  "model": "ollama/qwen3.6-27b"
}

LM Studio provider block

jsonc

"options": {
  "baseURL": "http://127.0.0.1:1234/v1",
  "apiKey": "lm-studio"
}

Restart OpenCode after editing config, or run opencode fresh from the project directory.

Step 5 — Cloud + local in one config (recommended)

Post-Fable tiering:

jsonc

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "llama": {
      "name": "Local Qwen",
      "npm": "@ai-sdk/openai-compatible",
      "options": {
        "baseURL": "http://127.0.0.1:8080/v1",
        "apiKey": "local"
      },
      "models": {
        "qwen3.6-27b": { "name": "Qwen3.6-27B local" }
      }
    },
    "zai": {
      "name": "GLM Coding Plan",
      "options": {
        "baseURL": "https://api.z.ai/api/coding/paas/v4"
      }
    }
  },
  "model": "llama/qwen3.6-27b"
}

Workload	Switch to
Private repo, offline flight	`llama/qwen3.6-27b`
Hard refactor, long agent	`/models` → GLM-5.2 (harness guide)
Cheap volume API	Cline $9.99 open weights or OpenRouter

Connect Z.AI via /connect inside TUI if you prefer not to store API keys in jsonc.

Model picks for local OpenCode (July 2026)

Model	Local fit	OpenCode notes
Qwen 3.6 27B dense	Best balance 48GB Mac / 24GB+ Nvidia	Beats MoE 35B A3B on instruction following
Gemma 4 12B/31B	Multimodal + Apache 2.0	Slower than Qwen for pure code
GLM-5.2 Unsloth	Workstation / multi-GPU	Closest to frontier local
DeepSeek V4 Flash (quant)	High tok/s if RAM allows	Aggressive quants trade quality
Kimi K2.7	Usually API — too large for most locals	Use API via OpenCode `/connect`

Alibaba cloud line: Qwen 3.7-Max for when local is not enough.

Hardware reality check

Machine	Realistic local model
32GB Apple Silicon	Q4 27B or Q8 7B–14B
48–64GB Apple Silicon	Q8 Qwen 3.6 27B (~32 tok/s MTP)
24GB Nvidia (4090/5090)	Q6 27B; Q4 70B tight
64GB+ VRAM / dual GPU	GLM-5.2 class, vLLM team server

Full economics: Mac vs dedicated GPU · closed vs open cost table.

First session checklist

Server running — curl -s http://127.0.0.1:8080/v1/models (or Ollama 11434)
Config saved — ~/.config/opencode/opencode.jsonc
opencode in git repo
/init — creates AGENTS.md
Smoke prompt — "Create a pnpm package with a hexagonal minesweeper" (same test as Quesma Qwen post)
Verify model — ask "What model are you?" — should not hallucinate Claude/GPT

Add verification loops from explainx.ai loops — e.g. ci-until-green.

Troubleshooting

Symptom	Fix
Connection refused	Inference server not running or wrong port in baseURL
Empty / garbage output	Quant too low — bump Q4→Q6/Q8 (quant guide)
Slow first token	Model loading; MTP helps decode — see llama.cpp `--spec-type draft-mtp`
Ignores package.json / structure	Try dense over MoE; shorten context if RAM swapping
Tool calls fail	Model may lack reliable function-calling — switch to GLM API or GPT-class for agent-heavy runs
Wrong provider	`echo` config path; `/models` list; check `model` string matches `provider/id`

OpenCode vs other local harnesses

Harness	Local OSS models	Notes
OpenCode	Yes — 75+ providers + custom local	Default open multi-surface
Pi	Yes — provider plugins	Minimal, ownable
Codex OSS	Yes — `--oss` + Ollama	OpenAI-native tool schema
Claude Code	Indirect — `/config` remote to desktop	Not weight-local
Kilo Code	Yes — BYOK + Ollama	VS Code extension

OpenCode wins when you want one harness, terminal + desktop, and swap local/cloud without reinstalling.

Official: OpenCode · OpenCode config schema · llama.cpp · Ollama

Runtime ports, model tags, and OpenCode provider schema reflect July 2026 docs — verify before production. Last updated: July 17, 2026.

This is explainx.ai's end-to-end stack guide: pick a model → start a server → configure OpenCode → tier local vs cloud.

TL;DR — the full local + OpenCode path

Step	What to do	Deep dive
1. Hardware	32GB+ for Q4 7B–27B; 48GB+ for Q8 27B	Mac vs GPU guide
2. Model	Qwen 3.6 27B coding default; GLM-5.2 if you have GPU headroom	Qwen local guide · GLM Unsloth
3. Runtime	llama.cpp (control) or Ollama (ease)	What is llama.cpp? · Personal AI system
4. Server	OpenAI-compatible `/v1` on localhost	Sections below
5. OpenCode	`opencode.jsonc` provider + `opencode` in repo	OpenCode harness guide
6. Verify	`/init`, small coding task, check tok/s	Quantization guide

Architecture — three layers

text

┌─────────────────────────────────────┐
│  OpenCode (agent harness)           │
│  tools · LSP · AGENTS.md · sessions │
└──────────────┬──────────────────────┘
               │ OpenAI-compatible HTTP
┌──────────────▼──────────────────────┐
│  Inference server                   │
│  llama.cpp · Ollama · LM Studio · vLLM│
└──────────────┬──────────────────────┘
               │ GGUF / safetensors
┌──────────────▼──────────────────────┐
│  Open-weight model on your disk/GPU │
│  Qwen · GLM · Gemma · DeepSeek · …  │
└─────────────────────────────────────┘

OpenCode never loads weights itself — it only talks HTTP. That separation is why one config file can swap Ollama today and llama.cpp tomorrow.

Step 1 — Install OpenCode

bash

curl -fsSL https://opencode.ai/install | bash
# or: npm install -g opencode-ai
# or: brew install anomalyco/tap/opencode

bash

cd your-project
opencode

First-run inside TUI:

/connect — for cloud providers (Z.AI, OpenRouter, Copilot OAuth)
/init — generate AGENTS.md project memory
/models — pick active model after providers exist

Slash reference: OpenCode commands. Harness concepts: What is an agent harness?.

Step 2 — Pick an inference runtime

Runtime	Best for	OpenAI API default	OpenCode fit
llama.cpp	Max control, MTP, Apple Silicon	`http://127.0.0.1:8080/v1`	Recommended for serious local coding
Ollama	One-command pull/serve	`http://127.0.0.1:11434/v1`	Fastest onboarding
LM Studio	GUI + local server toggle	`http://127.0.0.1:1234/v1`	Non-terminal users
vLLM	Multi-user / production GPU box	custom port `/v1`	Team LAN server

Codex users: same local servers work in Codex OSS mode — different harness, identical inference layer.

Step 3 — Start the local server

Option A — llama.cpp (Qwen 3.6 27B example)

From the Qwen 3.6 local benchmark post:

bash

llama-server -hf unsloth/Qwen3.6-27B-MTP-GGUF:Q8_0 \
    --spec-type draft-mtp -ngl 999 -fa on -c 65536 --port 8080

Smoke test: open http://127.0.0.1:8080 or curl http://127.0.0.1:8080/v1/models.

Option B — Ollama

bash

ollama pull qwen3.6:27b        # verify tag on ollama.com/library
ollama serve                   # usually auto-starts on 11434

Other coding pulls teams use: glm-5.2, deepseek-coder-v2, gemma4:12b — check VRAM before pulling 70B-class tags.

Option C — LM Studio

Download model in GUI
Local Server tab → Start Server (default 1234)
Enable OpenAI compatible API

Option D — vLLM (Linux GPU server)

bash

python -m vllm.entrypoints.openai.api_server \
  --model Qwen/Qwen3.6-27B-Instruct \
  --port 8000

Point OpenCode at http://127.0.0.1:8000/v1 or your LAN IP for a shared team box.

Step 4 — Configure OpenCode (`opencode.jsonc`)

Config path: ~/.config/opencode/opencode.jsonc (user-wide) or project-local per OpenCode docs.

llama.cpp provider block

jsonc

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "llama": {
      "name": "llama.cpp (local)",
      "npm": "@ai-sdk/openai-compatible",
      "options": {
        "baseURL": "http://127.0.0.1:8080/v1",
        "apiKey": "local"
      },
      "models": {
        "qwen3.6-27b": { "name": "Qwen3.6-27B Q8 +MTP" }
      }
    }
  },
  "model": "llama/qwen3.6-27b"
}

Ollama provider block

jsonc

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "ollama": {
      "name": "Ollama (local)",
      "npm": "@ai-sdk/openai-compatible",
      "options": {
        "baseURL": "http://127.0.0.1:11434/v1",
        "apiKey": "ollama"
      },
      "models": {
        "qwen3.6-27b": { "name": "qwen3.6:27b" }
      }
    }
  },
  "model": "ollama/qwen3.6-27b"
}

LM Studio provider block

jsonc

"options": {
  "baseURL": "http://127.0.0.1:1234/v1",
  "apiKey": "lm-studio"
}

Restart OpenCode after editing config, or run opencode fresh from the project directory.

Step 5 — Cloud + local in one config (recommended)

Post-Fable tiering:

jsonc

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "llama": {
      "name": "Local Qwen",
      "npm": "@ai-sdk/openai-compatible",
      "options": {
        "baseURL": "http://127.0.0.1:8080/v1",
        "apiKey": "local"
      },
      "models": {
        "qwen3.6-27b": { "name": "Qwen3.6-27B local" }
      }
    },
    "zai": {
      "name": "GLM Coding Plan",
      "options": {
        "baseURL": "https://api.z.ai/api/coding/paas/v4"
      }
    }
  },
  "model": "llama/qwen3.6-27b"
}

Workload	Switch to
Private repo, offline flight	`llama/qwen3.6-27b`
Hard refactor, long agent	`/models` → GLM-5.2 (harness guide)
Cheap volume API	Cline $9.99 open weights or OpenRouter

Connect Z.AI via /connect inside TUI if you prefer not to store API keys in jsonc.

Model picks for local OpenCode (July 2026)

Model	Local fit	OpenCode notes
Qwen 3.6 27B dense	Best balance 48GB Mac / 24GB+ Nvidia	Beats MoE 35B A3B on instruction following
Gemma 4 12B/31B	Multimodal + Apache 2.0	Slower than Qwen for pure code
GLM-5.2 Unsloth	Workstation / multi-GPU	Closest to frontier local
DeepSeek V4 Flash (quant)	High tok/s if RAM allows	Aggressive quants trade quality
Kimi K2.7	Usually API — too large for most locals	Use API via OpenCode `/connect`

Alibaba cloud line: Qwen 3.7-Max for when local is not enough.

Hardware reality check

Machine	Realistic local model
32GB Apple Silicon	Q4 27B or Q8 7B–14B
48–64GB Apple Silicon	Q8 Qwen 3.6 27B (~32 tok/s MTP)
24GB Nvidia (4090/5090)	Q6 27B; Q4 70B tight
64GB+ VRAM / dual GPU	GLM-5.2 class, vLLM team server

Full economics: Mac vs dedicated GPU · closed vs open cost table.

First session checklist

Server running — curl -s http://127.0.0.1:8080/v1/models (or Ollama 11434)
Config saved — ~/.config/opencode/opencode.jsonc
opencode in git repo
/init — creates AGENTS.md
Smoke prompt — "Create a pnpm package with a hexagonal minesweeper" (same test as Quesma Qwen post)
Verify model — ask "What model are you?" — should not hallucinate Claude/GPT

Add verification loops from explainx.ai loops — e.g. ci-until-green.

Troubleshooting

Symptom	Fix
Connection refused	Inference server not running or wrong port in baseURL
Empty / garbage output	Quant too low — bump Q4→Q6/Q8 (quant guide)
Slow first token	Model loading; MTP helps decode — see llama.cpp `--spec-type draft-mtp`
Ignores package.json / structure	Try dense over MoE; shorten context if RAM swapping
Tool calls fail	Model may lack reliable function-calling — switch to GLM API or GPT-class for agent-heavy runs
Wrong provider	`echo` config path; `/models` list; check `model` string matches `provider/id`

OpenCode vs other local harnesses

Harness	Local OSS models	Notes
OpenCode	Yes — 75+ providers + custom local	Default open multi-surface
Pi	Yes — provider plugins	Minimal, ownable
Codex OSS	Yes — `--oss` + Ollama	OpenAI-native tool schema
Claude Code	Indirect — `/config` remote to desktop	Not weight-local
Kilo Code	Yes — BYOK + Ollama	VS Code extension

OpenCode wins when you want one harness, terminal + desktop, and swap local/cloud without reinstalling.

Official: OpenCode · OpenCode config schema · llama.cpp · Ollama

Runtime ports, model tags, and OpenCode provider schema reflect July 2026 docs — verify before production. Last updated: July 17, 2026.

TL;DR — the full local + OpenCode path

Architecture — three layers

Step 1 — Install OpenCode

Step 2 — Pick an inference runtime

Step 3 — Start the local server

Option A — llama.cpp (Qwen 3.6 27B example)

Option B — Ollama

Option C — LM Studio

Option D — vLLM (Linux GPU server)

Step 4 — Configure OpenCode (opencode.jsonc)

llama.cpp provider block

Ollama provider block

LM Studio provider block

Step 5 — Cloud + local in one config (recommended)

Model picks for local OpenCode (July 2026)

Hardware reality check

First session checklist

Troubleshooting

OpenCode vs other local harnesses

Related on explainx.ai

TL;DR — the full local + OpenCode path

Architecture — three layers

Step 1 — Install OpenCode

Step 2 — Pick an inference runtime

Step 3 — Start the local server

Option A — llama.cpp (Qwen 3.6 27B example)

Option B — Ollama

Option C — LM Studio

Option D — vLLM (Linux GPU server)

Step 4 — Configure OpenCode (opencode.jsonc)

llama.cpp provider block

Ollama provider block

LM Studio provider block

Step 5 — Cloud + local in one config (recommended)

Model picks for local OpenCode (July 2026)

Hardware reality check

First session checklist

Troubleshooting

OpenCode vs other local harnesses

Related on explainx.ai

Related posts

Qwen 3.6 27B Local Dev Guide: llama.cpp, OpenCode, and Why Dense Beats MoE

LM Studio Bionic: Open-Model Agent for Code and Work Projects

What Is llama.cpp? Install, Run GGUF Models, and Serve OpenAI-Compatible APIs

Related posts

Qwen 3.6 27B Local Dev Guide: llama.cpp, OpenCode, and Why Dense Beats MoE

LM Studio Bionic: Open-Model Agent for Code and Work Projects

What Is llama.cpp? Install, Run GGUF Models, and Serve OpenAI-Compatible APIs

Step 4 — Configure OpenCode (`opencode.jsonc`)

Step 4 — Configure OpenCode (`opencode.jsonc`)