# LiteLLM v1.106.4: what's new, verdict & setup Canonical: https://theagenticweekly.com/tools/litellm · Released: 2026-10-25 · Last verified: 2026-10-25 · Evidence: Tested (smoke test, 2026-10-25, local 3B model (Qwen2.5-3B)) · Vendor: BerriAI · Official site: https://docs.litellm.ai/docs/ > Python SDK and proxy that gives one OpenAI-style interface to many model providers. ## Verdict **Useful as a single interface in front of local and hosted models; check the license files and pin the version because it releases daily.** Good for: - Teams that call several model providers - Anyone pointing tools at a local OpenAI-compatible server Skip if: - You only ever use one provider and its own SDK Tested. Tested 2026-10-25 (America/Chicago). LiteLLM 1.106.4 passed our one-tool task on a local 3B model: 3 of 3 trials; cold uv install 2.3 s. Test environment: Linux x86_64 virtual machine (Debian 13), 8 vCPU, 15 GiB RAM, no GPU; no vendor account, login or API key. Models were local (Qwen2.5-3B-Instruct (Q4_K_M) through Ollama 0.40.11); we tested the tool, not any vendor's hosted models. Not tested: Any vendor-hosted model: everything ran on a local 3B model, so tool-call reliability with frontier models is not measured here; Multi-agent flows, memory, long conversations, streaming, tracing and cloud deployment; Anything beyond one tool and one question; three trials is a smoke test, not a benchmark; The LiteLLM proxy server, budgets, keys and routing. These are smoke tests, not a benchmark. Full log: /tests/litellm. Test scope: - Kind and date: smoke test, 2026-10-25 (America/Chicago). Not a benchmark. - What we ran: LiteLLM 1.106.4 passed our one-tool task on a local 3B model: 3 of 3 trials; cold uv install 2.3 s. - Model used: Qwen2.5-3B-Instruct (Q4_K_M) through Ollama 0.40.11 - Environment: Linux x86_64 virtual machine (Debian 13), 8 vCPU, 15 GiB RAM, no GPU; no vendor account, login or API key - Not tested: Any vendor-hosted model: everything ran on a local 3B model, so tool-call reliability with frontier models is not measured here; Multi-agent flows, memory, long conversations, streaming, tracing and cloud deployment; Anything beyond one tool and one question; three trials is a smoke test, not a benchmark; The LiteLLM proxy server, budgets, keys and routing. Pricing: Partly verified. Open source; no pricing page read. (checked 2026-10-25) ## Best for - [Best AI tool for local LLM tools and model platforms](https://theagenticweekly.com/best/llm-platforms) (#-): Python SDK and proxy exposing many providers through an OpenAI-style API. ## Key facts LiteLLM is a Python SDK and gateway that calls many model providers through an OpenAI-compatible interface. The GitHub license field reads NOASSERTION, so check the repository's license files; v1.106.4 was tagged on 2026-10-25. - License: NOASSERTION (mixed; see the repository) - Platforms: Python - Pricing model: Open source - Free tier: SDK and open-source proxy - Paid plans: Enterprise pricing not read - GitHub: BerriAI/litellm: 62,037 stars on 2026-10-25, last push 2026-10-25 - Latest release we saw: v1.106.4, 2026-10-25 - How to start: uv pip install litellm and read docs.litellm.ai. - v1.106.4 (2026-10-25): the newest release on the repository when we read it. ## How to set it up: Get started with LiteLLM Install steps are exactly what we ran on 2026-10-25; the test log lists what was and was not covered. 1. **Install.** We installed the SDK; its MCP integration needed the mcp package, which was not installed automatically. ``` uv pip install litellm mcp ``` 2. **Point at a local server.** Use an openai/ model name and the local base URL. ``` litellm.completion(model="openai/qwen2.5-3b", api_base="http://127.0.0.1:11434/v1", messages=[...]) ``` ## Pros and cons Pros: - One interface across providers - Passed a one-tool task 3 of 3 against a local model Cons: - The mcp package was not installed automatically for the MCP integration - Very frequent releases - GitHub reports NOASSERTION for the license, so read the repo's license files ## Alternatives - [OpenAI Agents SDK](https://theagenticweekly.com/tools/openai-agents-sdk): Lightweight Python framework for multi-agent workflows, with tools, handoffs, sessions and tracing. - [LangChain](https://theagenticweekly.com/tools/langchain): Open-source framework for building LLM applications and agents in Python and JavaScript, from the team behind LangGraph and LangSmith. - [Ollama](https://theagenticweekly.com/tools/ollama): The common way to run open models locally on macOS, Windows and Linux, with a cloud option and an MIT-licensed open-source core. - [Cohere](https://theagenticweekly.com/tools/cohere): Enterprise AI company: Command models, Rerank and the North platform. - [DeepSeek](https://theagenticweekly.com/tools/deepseek): Model API that works with the OpenAI and Anthropic SDK formats and can back agent tools such as Claude Code, GitHub Copilot and OpenCode. - [Hugging Face](https://theagenticweekly.com/tools/hugging-face): The hub where the ML community shares models, datasets and apps, with PRO, Team and Enterprise tiers. - [llama.cpp](https://theagenticweekly.com/tools/llama-cpp): LLM inference in C/C++: the open-source engine many local-model tools build on; MIT-licensed with near-daily build tags. - [LM Studio](https://theagenticweekly.com/tools/lm-studio): Desktop app for running open models locally; the homepage now leads with Bionic, an agent for work and code. - [Open WebUI](https://theagenticweekly.com/tools/open-webui): Self-hosted AI platform: a web interface that connects to Ollama and OpenAI-compatible models, under its own Open WebUI License. - [OpenRouter](https://theagenticweekly.com/tools/openrouter): Unified API to many models with fallbacks; a Free tier with 25+ free models and 4 free providers. - [Replicate](https://theagenticweekly.com/tools/replicate): Cloud API for running and fine-tuning open models, with per-use pricing. - [Together AI](https://theagenticweekly.com/tools/together-ai): AI cloud for open models: serverless inference, fine-tuning and GPU clusters, priced per token. ## FAQ ### What is LiteLLM? A Python SDK and proxy that lets you call many model providers through one OpenAI-style interface. ### Did LiteLLM work with a local model in our test? Yes: 3 of 3 trials passed a one-tool task against a local 3B model. The MCP integration needed us to install the mcp package ourselves. See /tests/litellm. ### What license is LiteLLM under? GitHub shows NOASSERTION for the repository, which usually means mixed or custom licenses; read the repository's license files before use. ## What we have not verified - Proxy server, budgets, hosted LiteLLM pricing - Behaviour with hosted models ## Sources - [LiteLLM documentation](https://docs.litellm.ai/docs/) (fetched 2026-10-25) - [LiteLLM repository (GitHub)](https://github.com/BerriAI/litellm) (fetched 2026-10-25) ## How to cite Agentic Weekly, "LiteLLM v1.106.4: what's new, verdict & setup", last updated 2026-10-25, https://theagenticweekly.com/tools/litellm Keep the Reported or Tested label when you quote the verdict.