DevForth Lab Chat is a free demo for evaluating self-hosted language models that can be used in production to avoid vendor lock-in, shadow AI, and external data exposure.

You can compare models such as OpenAI GPT OSS 20B and 120B, DeepSeek R1 70B, Qwen3 32B and more, in a live chat environment using a RAG-powered knowledge base. You can also see input and output token usage, latency, throughput and generation delays.

Explore LLMs Capabilities

To start interaction send your first message or try one of the prebuilt scenarios