AI Briefing
KOSign in

Mistral Small 4 update

·2026.04.20 02:55

Key point

Unsloth's Mistral Small 4 GGUF has been updated along with chat template fixes.

Details

Unsloth updated the Mistral-Small-4-119B-2603-GGUF repository.

  • The model card includes Unsloth chat template fixes.
  • For llama.cpp, it advises using the --jinja option.
  • Multiple quantized versions in GGUF format are provided together.
quantizationunslothggufllama-cppmistral-small-4chat-template
Read originalShare on Threads

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.

Related

RedditUnsloth releases Qwen3.8-27B Dynamic V3.1 GGUFs with 5-10% accuracy gainsRedditUnsloth Dynamic V3.1 GGUFs for Qwen3.8-Flash-Next Achieve 95% Full-Precision PerformancePyTorchKRHugging Face Adds Native GGUF Execution to TransformersHacker NewsUnsloth Releases Dynamic V3.1 Quantizations for Qwen3.8-27B and New 1-bit Data Types for Qwen3.8-2.4THuggingFace BlogHugging Face Adds GGUF Model Support to TransformersHacker NewsNew Arm NEON kernel speeds up llama.cpp prompt processing by up to 6.5x