corpus.blog
41554 blogs · 4,417,066 posts Claim your blog
Record · corpus.blog/posts/01a0879a-e0df-70eb-93e3-4e4f327d0cf0

125B Qwen3.8-Flash-Next hits 18 tok/s on M1 Max 64GB with N-gram table on SSD

lilting.ch · published 27 August 2026

A record is what was published and who pointed at it. The text of the post is not held here: read it at the source, or at its Wayback capture.

Read the post at lilting.ch Wayback capture
0blogs citing
First cited
—
Most recent
—
Rank this month
not ranked
Links from this post
8
Words captured
4,173

Cited by

No blog we hold has cited this post yet.

Links from this post

Link
Host
Qwen3.8-Flash-Next 125B on EVO-X2 gfx1151: tg128 23.26, 128K context ceiling Read at the source ↗
lilting.ch
Qwen3.8-27B on M1 Max 64GB: MLX vs Ollama, and a 50k-char Thinking Runaway Read at the source ↗
lilting.ch
https://huggingface.co/Qwen/Qwen3.8-Flash-Next
huggingface.co
https://github.com/QwenLM/Qwen3.8-Flash-Next
github.com
https://github.com/ggml-org/llama.cpp/pull/27742
github.com
https://huggingface.co/AtomicChat/Qwen3.8-Flash-Next-GGUF
huggingface.co
https://unsloth.ai/docs/models/qwen3.8-next
unsloth.ai
Qwen3.8-27B on Strix Halo (EVO-X2): ROCm llama.cpp + MTP at 22 tok/s in Q8_0 Read at the source ↗
lilting.ch