server/v1.11.2

• parse thinking part for third-party models (qwen, deepseek reasoning styles)

server/v1.11.0

• third-party models configurable headers
• drop support for refact, starcoder2, deepseek-coder
• [experimental] refact proxy: lightweight version of server with third-party models proxying

server/v1.10.0

• Configurable Third-Party APIs: set up any chat model supported by litellm (available from JetBrains 6.3.2, VSCode 6.3.5)
• Model deprecation: starcoder2*, deepseek-coder*, refact

server/v1.9.3

Features:

• claude-3-7-sonnet, o3-mini models
• support of reasoning capabilities

v1.9.1

Features:

• cpu embeddings model thenlper/gte-base/cpu
• manual upload weights for offline mode

v1.9.0

Features:

• new 3rdparty providers: Gemini, Grok (XAI), OpenAI o1
• offline mode: server now works with preloaded models without connection to huggingface

Minor fixes:

• caps reload on model list change fix
• removed deprecated models

v1.8.0

Refact.ai Self-hosted:

• CUDA and cuDNN Version Update: 11.8.0 -> 12.4.1
• New models: llama3.1, llama3.2 and qwen2.5/coder families.
• New providers support: 3rd party APIs for groq and cerebras.
• Support for multiline code completion models.

v1.7.0

Refact.ai Self-hosted:

• New models: last OpenAI models is now available in Docker.
• Tool usage support for 3rd party models: turn on 3rd party APIs to use latest features of Refact.
• Removed deprecated models.

v1.6.4

Refact.ai Self-hosted:

• Claude-3.5 Sonnet Support: New model from Anthropic is now available in Docker.

Refact.ai Enterprise:

• Llama3 vLLM Support: Added vLLM version of Llama-3-8B-Instruct for better performance.

v1.6.3

Refact.ai Self-hosted:

• Llama3 8k Context: llama3 models now support 8k context.
• Credentials Management: We added information about tokens and keys.
• Deprecated Models: The models starcoder, wizardlm, and llama2 are deprecated and will be removed in the next release.

Refact.ai Enterprise:

• Refact Model 4k Context: refact model now supports 4k context.