Skip to content

Releases: lofcz/LlmTornado

Release v3.4.4

17 Mar 00:00
82b55d3
Compare
Choose a tag to compare

What's Changed

This release adds useful abstractions to Conversations and Files: WaitForContentReady, and WaitForReady provide a concurrent polling primitive. Two new callbacks are available for streaming responses: MessagePartHandler, ImageTokenHandler. New models supported: gemma3-27b, gemini-2.0-flash-exp-image-generation

demo

  • link yt demo in readme
  • google: support multimodal outputs
  • conversation: WaitForContentReady, files: WaitForReady, openai: automatically include usage when streaming
  • google: gemma3-27b
  • Update project versions to v3.4.3

Updated Projects

LlmTornado.csproj -> 3.4.4
LlmTornado.Demo.csproj -> 0.0.1

Release v3.4.3

15 Mar 22:57
Compare
Choose a tag to compare

What's Changed

This release adds support for command-a-03-2025.

  • fix
  • up tags
  • cohere: command-a-03-2025
  • Update README.md
  • Update FeatureMatrix.md

Updated Projects

LlmTornado.csproj -> 3.4.3
LlmTornado.Demo.csproj -> 0.0.1

Release v3.4.2

15 Mar 21:05
Compare
Choose a tag to compare

What's Changed

This release adds support for Mistral! Following models are supported: codestral-latest, mistral-large-latest, pixtral-large-latest, mistral-saba-latest, ministral-3b-latest, ministral-8b-latest, mistral-small-latest, pixtral-12b-2409, open-mistral-nemo, open-codestral-mamba.

  • mistral chat extensions
  • add ocr to the feature matrix
  • bootstrap mistral
  • map mistral models
  • Update FeatureMatrix.md
  • Update project versions to v3.4.1

Updated Projects

LlmTornado.csproj -> 3.4.2
LlmTornado.Demo.csproj -> 0.0.1

Release v3.4.1

15 Mar 18:17
Compare
Choose a tag to compare

What's Changed

This release adds support for Google Embeddings: text-embedding-004 is now supported.

  • google: support embeddings
  • Update FeatureMatrix.md
  • Update README.md
  • Update project versions to v3.4.0

Updated Projects

LlmTornado.csproj -> 3.4.1
LlmTornado.Demo.csproj -> 0.0.1

Release v3.4.0

15 Mar 16:23
18c50ca
Compare
Choose a tag to compare

What's Changed

This release adds support for DeepSeek! Models deepseek-chat and deepseek-reasoner are now supported, including streaming, tool calls, JSON schema, any many more features! 🎉
OpenAI reasoning models now correctly ignore temperature when sterilizing chat requests.

  • include deepseek in readme
  • mark deepseek as supported
  • deepseek: deepseek-chat, deepseek-reasoner, openai: fix tempereature for reasoning models
  • Update project versions to v3.3.2

Updated Projects

LlmTornado.csproj -> 3.4.0
LlmTornado.Demo.csproj -> 0.1.0

Release v3.3.2

09 Mar 13:43
17c2aaf
Compare
Choose a tag to compare

What's Changed

This release implements the missing createThreadAndRun endpoint from OpenAI's Assistants.

  • code style adjustments
  • Adding support for creating thread and run in one request
  • add test for non-streaming file upload
  • Update project versions to v3.3.1

Updated Projects

LlmTornado.csproj -> 3.3.2
LlmTornado.Demo.csproj -> 0.0.1

Release v3.3.1

07 Mar 06:42
Compare
Choose a tag to compare

What's Changed

This release adds support for Cohere's c4ai-aya-vision-32b. In the next minor release, we will migrate to the v2 API to support the vision modality.

  • cohere: c4ai-aya-vision-32b, c4ai-aya-vision-8b
  • up tests
  • Update project versions to v3.3.0

Updated Projects

LlmTornado.csproj -> 3.3.1
LlmTornado.Demo.csproj -> 0.0.1
LlmTornado.Tests.csproj -> 0.0.1

Release v3.3.0

02 Mar 10:39
b0874d7
Compare
Choose a tag to compare

What's Changed

This minor release refactors Image Generation APIs to bring them on par with other endpoints. Several image option classes are now enums instead. imagen-3.0-generate-002 is now supported.

dalle3
imagen3

  • Update FeatureMatrix.md
  • google: imagen-3.0-generate-002
  • Update project versions to v3.2.8

Updated Projects

LlmTornado.csproj -> 3.3.0
LlmTornado.Demo.csproj -> 0.1.0

Release v3.2.8

28 Feb 04:00
Compare
Choose a tag to compare

What's Changed

This release adds support for the following models: gpt-4.5-preview, command-r7b-arabic-02-2025, c4ai-aya-expanse-32b. Image over URL is now supported for Anthropic, documents both as base64 and URL too. ToolChoice = none is now supported for Anthropic. ToolChoice static helpers were added.

  • anthropic: support document message parts
  • anthropic: automatically strip base64 header from imgs
  • anthropic: support url images
  • cohere: c4ai-aya-expanse-8b, c4ai-aya-expanse-32b
  • cohere: command-r7b-arabic-02-2025
  • anthropic: support newly added "none" tool choice
  • openai: gpt-4.5-preview, gpt-4.5-preview-2025-02-27
  • chore: update EndpointBase.cs
  • Update project versions to v3.2.7

Updated Projects

LlmTornado.csproj -> 3.2.8
LlmTornado.Demo.csproj -> 0.0.1

Release v3.2.7

26 Feb 04:31
Compare
Choose a tag to compare

What's Changed

This release adds support for gemini-2.0-flash-lite.

  • google: gemini-2.0-flash-lite
  • Update README.md
  • Update README.md
  • Update README.md
  • improve demo repl
  • Update project versions to v3.2.6

Updated Projects

LlmTornado.csproj -> 3.2.7
LlmTornado.Demo.csproj -> 0.0.1