Qwen3.6-Whittle-25B-A3B
logic65基础档案已收录,详细介绍后续补充。可先查看厂商、分类与来源。
从语言到影像,从开源到 API。读懂模型能力、比较真实成本,
为你的下一个想法,找到合适的智能引擎。
基础档案已收录,详细介绍后续补充。可先查看厂商、分类与来源。
> ### ☕ Support this work > Whittle is built by one person on a grocery budget and rented GPU hours. If this research is useful to you, or you want to see it finished: > **ko-fi.com/davida81328**. Every hour of GPU time goes straight into the next checkpoint, and every checkpoint, table and log lands in these repos. **A 27B whittled down to 16.8B with a logit lens and a pricing table, then healed with one A100 evening. It now outscores every intermediate cut, including ones 4B larger, at 20 tokens/second on two consumer 8GB GPUs.** Whittle is Qwen3.8-27B minus 20 of its 64 layers (whole interval-blocks, chosen by measured boundary cost) and minus 25% of every remaining MLP's width (weakest neurons by ‖down_column‖ × activation-std). That is a −10.1B parameter cut performed with **zero training**, followed by a single 11M-token QLoRA heal on a fully clean-lineage mix (synthetic fact frames, programmatic short-think arithmetic, hand-written code drills, public-domain long pages).
译文排队中 · 当前显示来源原文
> ### ☕ Support this work > Whittle is built by one person on a grocery budget and rented GPU hours. If this research is useful to you, or you want to see it finished: > **ko-fi.com/davida81328**. Every hour of GPU time goes straight into the next checkpoint, and every checkpoint, table and log lands in these repos. Quantized builds of Whittle MoE 27B (A18B), a post hoc mixture of experts carved from Qwen3.8-27B and taught when to stop talking. v2.1 passed every release bar: 8 percent single turn loop rate (down from 69 at first release), 22 percent structured (down from 75), zero truncated answers, knowledge battery 28 of 39. Method, measurements, and the failures along the way are on the main model card. Earlier line (Aug 2026); the current Whittle models are the Whittle-Next / Whittle-Qwen-3.8 line: Whittle-Qwen-3.8-35B-A3B.
译文排队中 · 当前显示来源原文
基础档案已收录,详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
基础档案已收录,详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
基础档案已收录,详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
基础档案已收录,详细介绍后续补充。可先查看厂商、分类与来源。
基础档案已收录,详细介绍后续补充。可先查看厂商、分类与来源。
基础档案已收录,详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
基础档案已收录,详细介绍后续补充。可先查看厂商、分类与来源。
> ### ☕ Support this work > Whittle is built by one person on a grocery budget and rented GPU hours. If this research is useful to you, or you want to see it finished: > **ko-fi.com/davida81328**. Every hour of GPU time goes straight into the next checkpoint, and every checkpoint, table and log lands in these repos. Ready-to-run **llama.cpp** quantisations of logic65/Whittle-Qwen-3.8-35B-A3B: a 35.1 B-total, ~3 B-active mixture-of-experts in the Qwen3.8-Flash-Next (`qwen4_exp`) format with a **10 B-parameter n-gram memory that is load-bearing** (the body measurably needs it; see the base card). Distilled from Qwen3.8-27B at the logit level on 25k verified teacher traces; built on Whittle-Next-27B-A3B. **Runs on stock llama.cpp, no patches.** The full card, measurements and caveats are on the base repo; this repo is the download. **Current.** These files are the **Phase-2 step 32010 root** (*p2-s32010*, 30 Sep 2026) of the current Whittle flagship: the complete Phase-2 distillation from Qwen3.8-27B (every usable teacher row seen once), a memory phase, and a 410-step fix for the llama.cpp web page's browser tools. The previous (p2-s6000) ladder was replaced in place; the earlier full weights stay under `bf16-s6000/`, `bf16-agentfix2/`, `bf16-lw5/`, `bf16-lw2/` and `bf16-tbl1/` on the base repo.
译文排队中 · 当前显示来源原文
基础档案已收录,详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
基础档案已收录,详细介绍后续补充。可先查看厂商、分类与来源。
基础档案已收录,详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
可以调整厂商、分类或输入输出条件;新模型正持续收录。
调整月度 Token 用量,直观比较不同模型的预估 API 开支。
使用已采集且单位可比的 USD / 百万 Token 渠道报价。报价为采集时快照;不计缓存优惠、阶梯价、推理额外输出与税费,请以来源的当前报价为准。
先区分输入、输出和缓存价格,再使用同一任务和用量估算总费用。图像、音频和视频可能按张、时长或分辨率计费,不能直接与 Token 单价比较。
不一定。必须逐一核对具体模型版本的许可证、商用限制及使用政策。自己部署还需要考虑显存、推理资源和运维费用。
榜单取决于测试任务、版本与评测方法。建议用实际业务样本比较准确性、时延、稳定性和成本,再选择模型。