The AI Toolchain — issue -224, January 5, 2026

Glamourous agentic coding for all

Crush v0.30.3 adds five new models for GitHub Copilot users: Claude Opus 4.5, GPT 5.2, GPT 5.1, GPT 5.1 Codex, and GPT 5.1 Max.

$ git clone --branch v0.30.3 https://github.com/charmbracelet/crush.git
# already have the repo? check out this version:
$ git checkout v0.30.3

• ›Adds Claude Opus 4.5 as a selectable model for GitHub Copilot users.
• ›Adds GPT 5.2, GPT 5.1, GPT 5.1 Codex, and GPT 5.1 Max as selectable models for GitHub Copilot users.

The open source coding agent.

OpenCode v1.1.3 adds Bedrock auth config, Japan cross-region inference, image-to-custom-commands, and CLI install in desktop.

$ git clone --branch v1.1.3 https://github.com/sst/opencode.git
# already have the repo? check out this version:
$ git checkout v1.1.3

• ›Supports jp. prefix for Amazon Bedrock cross-region inference profiles, enabling Japan-region routing.
• ›Adds configuration options and authentication precedence for Amazon Bedrock.
• ›Passes image parts to custom commands, enabling vision-capable custom command workflows.
• ›Adds system theme resolution and event handling to TUI.
• ›TUI autocomplete now expands directories on Tab and selects on Enter, improving filesystem navigation.

$ git clone --branch v1.1.2 https://github.com/sst/opencode.git
# already have the repo? check out this version:
$ git checkout v1.1.2

AI agent toolkit: unified LLM API, agent loop, TUI, coding agent CLI

Pi v0.37.0 adds headless OAuth over SSH, per-message shareable URLs, and a claude-rules extension example.

$ git clone --branch v0.37.0 https://github.com/earendil-works/pi.git
# already have the repo? check out this version:
$ git checkout v0.37.0

• ›Adds headless OAuth login with paste-based URL/code entry, enabling authentication over SSH sessions without a DISPLAY environment variable.
• ›Adds copy-link button in the share viewer to generate URLs that navigate directly to a specific message.
• ›Adds claude-rules extension example that loads .claude/rules/ entries into the system prompt.
• ›OpenAI Codex provider now sets thinking level separately from model selection, with the provider clamping to each model's supported range.

$ git clone --branch v0.36.0 https://github.com/earendil-works/pi.git
# already have the repo? check out this version:
$ git checkout v0.36.0

$ /login openai-codex

$ git clone --branch v0.35.0 https://github.com/earendil-works/pi.git
# already have the repo? check out this version:
$ git checkout v0.35.0

$ pi --extension ./safety.ts -e ./todo.ts

{
  "name": "my-extension-package",
  "dependencies": { "zod": "^3.0.0" },
  "pi": {
    "extensions": ["./src/main.ts", "./src/tools.ts"]
  }
}

Qwen Code v0.6.0-nightly adds locale-based LLM language auto-detection and ChatCompletionContentPart merging support.

$ git clone --branch v0.6.0-nightly.20260105.105ad743 https://github.com/QwenLM/qwen-code.git
# already have the repo? check out this version:
$ git checkout v0.6.0-nightly.20260105.105ad743

• ›Auto-detects the LLM output language from the system locale, eliminating the need to manually configure response language.
• ›Supports merging ChatCompletionContentPart objects and adds filtering of empty messages in API content handling.

Your own personal AI assistant. Any OS.

OpenClaw v2026.1.5 adds image-model support with a dedicated image tool and named model shorthands for major providers.

$ git clone --branch v2026.1.5 https://github.com/openclaw/openclaw.git
# already have the repo? check out this version:
$ git checkout v2026.1.5

agent:
  model: gpt-mini
  imageModel: gpt

agent:
  model: gemini-flash

• ›New agent.imageModel config key enables a dedicated image-specific model with fallback chain for image-aware scans.
• ›New image agent tool routes image tasks to the configured image model.
• ›Adds built-in model shorthands (opus, sonnet, gpt, gpt-mini, gemini, gemini-flash) for faster config authoring; user config takes precedence.

llama.cpp b7639 adds support for the LFM2-ColBERT-350M embedding model.

$ git clone --branch b7639 https://github.com/ggml-org/llama.cpp.git
# already have the repo? check out this version:
$ git checkout b7639

• ›Adds llama_model_n_embd_out() API function that returns hparams.n_embd_out when set, falling back to hparams.n_embd.
• ›Adds support for the LFM2-ColBERT-350M model architecture.

$ git clone --branch b7630 https://github.com/ggml-org/llama.cpp.git
# already have the repo? check out this version:
$ git checkout b7630