Unsloth has released updates for DeepSeek-V4.1 and FLUX.3, enabling local execution of these models. The DeepSeek-V4.1-Flash-1017 model boasts 552 billion parameters with 16 billion active and a 1 million token context window, running efficiently on SSDs. FLUX.3 Image Klein, a 5.6 billion parameter model, supports image editing with up to ten reference images and operates on 24GB of GPU memory. The release also introduces enhanced chat capabilities with image and video generation, multi-GPU training support via FSDP2, and a simplified MCP setup for coding agents. AI
IMPACT Enhances local model execution and training capabilities for developers using DeepSeek and FLUX models.
RANK_REASON This is a software release from Unsloth, a company that optimizes models, rather than a frontier model release from a primary lab.
- Black Forest Labs
- Claude Code
- Cursor+
- DeepSeek-V4.1-Flash-1017
- FLUX 3 Image Klein
- FSDP2
- MCP
- Qwen-Image-Edit-2509
- Qwen-Image-Edit-2511
- Unsloth
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →