Skip to content

This release adds 3 notable features for engineering teams evaluating rollout.

Published 15d LLM Frameworks
✓ No known CVEs patched
Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

ai apple-silicon benchmarks cli gguf gpu
+7 more
huggingface inference llm local-llm ollama python vram

Summary

AI summary

Fix crashes when running large Transformer models with an offload_folder.

Changes in this release

Feature Medium

Detect DGX Spark / NVIDIA GB10 as shared-memory NVIDIA GPU when memory.total unavailable.

Detect DGX Spark / NVIDIA GB10 as shared-memory NVIDIA GPU when memory.total unavailable.

Source: granite4.1:8b-q6_K@2026-05-19

Confidence: high

Feature Medium

Respect XDG_CACHE_HOME for cache paths, ignoring relative values per XDG spec.

Respect XDG_CACHE_HOME for cache paths, ignoring relative values per XDG spec.

Source: granite4.1:8b-q6_K@2026-05-19

Confidence: high

Feature Medium

Treat Apple Silicon as shared memory in fit detection.

Treat Apple Silicon as shared memory in fit detection.

Source: granite4.1:8b-q6_K@2026-05-19

Confidence: high

Performance Medium

Inline LiveBench fallback data and speed up benchmark score fetching.

Inline LiveBench fallback data and speed up benchmark score fetching.

Source: granite4.1:8b-q6_K@2026-05-19

Confidence: high

Bugfix Medium

`whichllm run` no longer crashes for large Transformers models with offload_folder.

`whichllm run` no longer crashes for large Transformers models with offload_folder.

Source: granite4.1:8b-q6_K@2026-05-19

Confidence: high

Full changelog

What's Changed

  • Detect DGX Spark / NVIDIA GB10 as a shared-memory NVIDIA GPU when NVIDIA reports memory.total as unavailable.
  • Fix whichllm run crashes for large Transformers models by providing an offload_folder.
  • Respect XDG_CACHE_HOME for cache paths, while ignoring relative values per the XDG spec.
  • Treat Apple Silicon as shared memory in fit detection.
  • Inline LiveBench fallback data and speed up benchmark score fetching.

Validation

  • ruff format --check .
  • ruff check .
  • pytest -q -s
  • python -m build
  • twine check dist/*

Weekly OSS security release digest.

The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.

No spam, unsubscribe anytime.

Share this release

Track Find the best local LLM for your hardware, ranked by benchmarks

Get notified when new releases ship.

Sign up free

About Find the best local LLM for your hardware, ranked by benchmarks

All releases →

Beta — feedback welcome: [email protected]