Swift 1.5 Flash Next Heretic quants add IQ4_XS and IQ3_S tuned for reverse engineering
jjusko20 · reddit · 2026-10-31
LocalLLaMA user trubisky released additional GGUF quants of Swift 1.5 Flash Next Heretic, an uncensored reverse-engineering edition (125B MoE, fits in 96GB VRAM):
- IQ4XS: based on the Unsloth recipe with 1GB of overhead where it mattered, cutting expert quantization error mass by 9.7%; 94.7GB on disk vs 109GB for the IQ4NL, same imatrix.
- IQ3S: the most requested format, a custom recipe with an 80/20 imatrix merge distribution (vs 70/30 for IQ4) to preserve reverse-engineering and reasoning ability at lower bit widths, trading some general knowledge; 85GB on disk (1.4GB over ISTA's for better accuracy).
All files are on Hugging Face at trubisky/Swift-1.5-Flash-Next-Heretic-ReverseEngineer.
More from Infra
- Clover-35B-A3B Quantized to 29.4 GB Runs at ~42 tok/s on a 64 GB Mac — viveky259 · 2026-10-31
- Liquid Agent brings fully local LFM3 agent to Linux with one-command install — JosephJacks_ · 2026-10-31
- A fully local office AI stack: MiniCPM, Qwen3, LFM2.5 and Whisper — HowDevelop · 2026-10-31
- a16z: GPU rental prices up 106% in 3 months while token prices fall 58% — a16z · 2026-10-31
- vLLM nightly adds PLE disk offload, with a telling [Qwen4Exp] tag on the PR — rmhubbert · 2026-10-31
- Solo AI studio: per-minute GPU rentals rendered a 328-frame Blender shot for ~$1 — r0ck3t23 · 2026-10-31