The AI Toolchain — issue -247, December 13, 2025

Get up and running with Kimi-K2.6, GLM-5.2, MiniMax, DeepSeek, gpt-oss, Qwen, Gemma and other models.

Ollama v0.13.4 adds Nemotron 3 Nano and OLMo 3/3.1 models, with Flash Attention now on by default.

$ git clone --branch v0.13.4 https://github.com/ollama/ollama.git
# already have the repo? check out this version:
$ git checkout v0.13.4

• ›Adds Nemotron 3 Nano, an efficient open model optimized for agentic workloads.
• ›Adds OLMo 3 and OLMo 3.1, open language models pre-trained on the Dolma 3 dataset and post-trained on Dolci datasets.
• ›Enables Flash Attention automatically by default for all models, improving inference performance.

Lightweight coding agent that runs in your terminal

Codex CLI v0.72.0 adds remote compaction for API-key sessions, expanded sandbox safety, and 1Password support.

$ git clone --branch rust-v0.72.0 https://github.com/openai/codex.git
# already have the repo? check out this version:
$ git checkout rust-v0.72.0

• ›Enables remote compacting (context compaction via API) for API-key-based sessions, not just OAuth users.
• ›Adds 1Password integration support.
• ›Introduces Elevated Sandbox levels 3 and 4, tightening sandboxing for shell command execution.
• ›Expands the safe-command allowlist, reducing unnecessary sandbox prompts for common operations.
• ›Adds reliable utilities for locating pwsh.exe and powershell.exe on Windows, and uses PowerShell to parse PowerShell syntax.

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

Pi v0.21.0 adds inline image rendering for Kitty/iTerm2 terminals and Gemini 3 Pro thinking-level support.

$ git clone --branch v0.21.0 https://github.com/earendil-works/pi.git
# already have the repo? check out this version:
$ git checkout v0.21.0

$ /show-images

• ›Renders images inline in tool output on terminals supporting Kitty graphics protocol (Kitty, Ghostty, WezTerm) or iTerm2 inline images, with aspect ratio preserved; toggle via /show-images command or terminal.showImages setting.
• ›Extends thinking level selector (Minimal/Low → Google LOW, Medium/High → Google HIGH) to Gemini 3 Pro models.

Qwen Code v0.5.0 adds Russian UI, terminal bell notifications, session resume hints, and bundles the CLI into the VSCode extension.

$ git clone --branch v0.5.0 https://github.com/QwenLM/qwen-code.git
# already have the repo? check out this version:
$ git checkout v0.5.0

• ›Adds a terminal bell setting to enable or disable audio notifications for completed operations.
• ›Shows the session resume command on exit so users can return to a previous conversation.
• ›Adds Russian language support to the UI.
• ›Bundles the CLI directly into the VSCode extension release package, removing the need for a separate CLI install.

$ git clone --branch v0.5.0-preview.0 https://github.com/QwenLM/qwen-code.git
# already have the repo? check out this version:
$ git checkout v0.5.0-preview.0

PydanticAI v1.32.0 adds tool timeouts, multi-agent Temporal workflow registration, and OTel log-based observability.

$ git clone --branch v1.32.0 https://github.com/pydantic/pydantic-ai.git
# already have the repo? check out this version:
$ git checkout v1.32.0

class MyWorkflow:
    __pydantic_ai_agents__ = [research_agent, summary_agent]

    async def run(self) -> str:
        ...

• ›Adds tool timeout support, allowing individual tools to be given a maximum execution duration.
• ›Allows TemporalAgents to be registered to a Temporal workflow via the __pydantic_ai_agents__ field on a workflow class.
• ›Extends end_strategy to apply to output tools in addition to regular tools, giving consistent early-exit behaviour across both tool types.
• ›Replaces OpenTelemetry events with OTel logs for agent observability, aligning with the OTel logging data model.

llama.cpp b7378 adds llama-completion to bash completion executables and warns of upcoming Linux archive format change.

$ git clone --branch b7378 https://github.com/ggml-org/llama.cpp.git
# already have the repo? check out this version:
$ git checkout b7378

• ›Adds llama-completion to the completion-bash executables, enabling tab-completion support for the llama-completion command in Bash.
• ›Linux releases will switch from .zip to .tar.gz archives in an upcoming release — deployment scripts should be updated accordingly.

$ git clone --branch b7376 https://github.com/ggml-org/llama.cpp.git
# already have the repo? check out this version:
$ git checkout b7376

$ git clone --branch b7366 https://github.com/ggml-org/llama.cpp.git
# already have the repo? check out this version:
$ git checkout b7366

$ llama-cli -mm ./mmproj-model.gguf -m ./model.gguf