Developments·Benchmark
Altworld's Hemmingway-1 (27B Qwen3.8 fine-tune) hits ~32 tok/s on M5 Max, excels at role-play fiction — llm-bench.io
A tester ran Altworld's Hemmingway-1, a 27B fine-tune of Qwen3.8-27B built for human-like writing, on a MacBook M5 Max at ~31.8 tok/s with thinking and MTP enabled. An LLM judge scored it 94.2 on role play and narrative, narrowly below one Qwen3.8-27B run (95.35), but side-by-side story samples read noticeably more like human fiction. The verdict: strong pick for fiction and everyday writing, while base Qwen remains the better generalist or coding assistant.
5 engagement1 sourceThu, Oct 1, 2026, 07:28 PM
AI Overview
- Altworld's Hemmingway-1-oQ8e-mtp is a 27B fine-tune of Qwen3.8-27B built for human-like writing, per llm-bench.io.
- On an M5 Max with thinking and MTP on, Hemmingway-1 averaged ~31.8 tok/s across 3 runs with ~29-31 GB VRAM peak.
- llm-bench.io's LLM judge scored Hemmingway-1 94.2 on role play/narrative, 88.2 research, 87.5 agent workflow, 76.2 code generation.
- Base Qwen3.8-27B-oQ8e-mtp scored close, with one run (95.35) beating all Hemmingway role-play runs (best 94.65).
- The reviewer finds Hemmingway-1's output reads like a natural short story while Qwen's feels 'technical', recommending Qwen for generalist or coding use.
Related
Releases·Model Release
Altworld open-sources Hemmingway-1, an Apache-2.0 27B creative-writing fine-tune of Qwen3.8-27B scoring EQ-Bench 4 1330
Altworld, a small lab based in Switzerland and South Africa, open-sourced Hemmingway-1, a 27B fine-tune of Qwen3.8-27B specialized for creative writing and short-form personal text. The Apache-2.0 weights are on Hugging Face and the model is vLLM-compatible. It posts EQ-Bench 4 1330 and tops the lab's internal blind pairwise writing benchmarks against the frontier models tested.
Altworld, a small lab based in Switzerland and South Africa, open-sourced Hemmingway-1, a 27B fine-tune of Qwen3.8-27B specialized for creative writing and short-form personal text. The Apache-2.0 weights are on Hugging Face and the model is vLLM-compatible. It posts EQ-Bench 4 1330 and tops the lab's internal blind pairwise writing benchmarks against the frontier models tested.
87 engagement2 sourcesredditSun, Sep 20, 2026, 07:54 PM
Social·Community
Swift-1.5-Qwen3.8-27b-oQ8e-mtp hits 34.8 tok/s on Apple M5 Max, averaging 51k tokens per llm-bench.io run vs Qwen3.8's 77k
A user ran Swift-1.5-Qwen3.8-27b-oQ8e-mtp through llm-bench.io on an Apple M5 Max (64 GB, oMLX, thinking at xhigh) and compared it to base Qwen3.8 27B. Swift 1.5 generates far fewer tokens, cutting full benchmark runs from ~38 to ~24 minutes, while quality judged as a tie (85.8 vs 85.0). The user plans to try it as a daily driver.
A user ran Swift-1.5-Qwen3.8-27b-oQ8e-mtp through llm-bench.io on an Apple M5 Max (64 GB, oMLX, thinking at xhigh) and compared it to base Qwen3.8 27B. Swift 1.5 generates far fewer tokens, cutting full benchmark runs from ~38 to ~24 minutes, while quality judged as a tie (85.8 vs 85.0). The user plans to try it as a daily driver.
17 engagement1 sourceredditTue, Sep 29, 2026, 04:07 PM
Social·Community
Qwen3.8-27B on M1 Max 32GB: MLX 15.8 tok/s vs llama.cpp 9.7 tok/s - but llama.cpp prefill is faster
A user benchmarked Qwen3.8-27B on an M1 Max Mac Studio with 32GB unified memory, comparing MLX and llama.cpp as local inference runtimes. MLX delivered higher decode throughput (15.8 tok/s vs 9.7 tok/s), while llama.cpp showed faster prefill. The results matter to practitioners choosing runtimes for memory-constrained Apple Silicon local LLM serving.
A user benchmarked Qwen3.8-27B on an M1 Max Mac Studio with 32GB unified memory, comparing MLX and llama.cpp as local inference runtimes. MLX delivered higher decode throughput (15.8 tok/s vs 9.7 tok/s), while llama.cpp showed faster prefill. The results matter to practitioners choosing runtimes for memory-constrained Apple Silicon local LLM serving.
4 engagement1 sourceredditMon, Sep 7, 2026, 06:36 PM
Products·launched
Qwen3.8-27B-Humanlike-Chat 2.0 — LoRA fine-tune making Qwen chat like a person, with working tool calls
A LoRA fine-tune of Qwen3.8-27B that makes the model converse in a casual, human-like tone instead of assistant speak. Version 2.0 fixes prior criticisms: functional tool calling, better instruction following, and more than a few words per reply. It targets local LLM users who dislike stiff assistant personalities.
A LoRA fine-tune of Qwen3.8-27B that makes the model converse in a casual, human-like tone instead of assistant speak. Version 2.0 fixes prior criticisms: functional tool calling, better instruction following, and more than a few words per reply. It targets local LLM users who dislike stiff assistant personalities.
933 engagement1 sourceredditFri, Oct 2, 2026, 04:00 PM
Releases·Tool Release
mlx-dspark v0.10.0 speeds Qwen3.8-27B up to 3× on Apple Silicon
mlx-dspark, an MLX port of DeepSeek's DSpark speculative-decoding drafters, now supports Qwen3.8-27B via RadixArk's drafter, achieving up to ~3× speedups on Apple Silicon. Benchmarks on an M4 Pro show 2.45× mean speedup for 8-bit targets, with code runs hitting 3.18×.
mlx-dspark, an MLX port of DeepSeek's DSpark speculative-decoding drafters, now supports Qwen3.8-27B via RadixArk's drafter, achieving up to ~3× speedups on Apple Silicon. Benchmarks on an M4 Pro show 2.45× mean speedup for 8-bit targets, with code runs hitting 3.18×.
210 engagement1 sourceredditFri, Aug 14, 2026, 09:54 PM