minor8
kinggeezero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。
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
Four ROCmFP4/ROCmFPX tiers of `IFM/K2-Horizon-MoVA-36B-A4B`, quantised from the official BF16 GGUF and measured on a Ryzen AI Max+ 395 (Radeon 8060S, gfx1151, 128 GB unified, ROCm 7.2.4). This is the **dual-expert** `k2_horizon` MoVA architecture — 36B total parameters, **~4B active** — with two independent expert systems: a 64-way MoVA value-expert bank (top-4) alongside a 100-way MoE feed-forward (top-8), 48 layers, hidden 2560, and it is a **reasoning** model. **Every number on this card was measured on these exact files.** Nothing is estimated. As an A4B MoE, it decodes far faster than a dense 32B on the same hardware. **Start with `FAST`** — fastest on both axes here, all four answer correctly. Take `AGENT` only if you specifically want 8-bit weights.
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。
`inclusionAI/Ling-3.0-flash-VL` (124B total / 5.5B active, native multimodal) quantized to ROCmFP4 for AMD Strix Halo, **with a multi-token-prediction head grafted in** from its text sibling — and with the server fix that lets **MTP and images work at the same time**, which they do not in stock llama.cpp. Everything below was measured on the shipped files on one Ryzen AI Max+ 395 (Radeon 8060S, 128 GiB unified). Numbers come from the built GGUFs and from server logs, never from estimates. ⚠️ **COHERENT is the worst of the three on this model** even though it is the largest and the highest bit-width (4.70 bpw vs 4.38). Its PPL is 0.80 worse than STRIX_LEAN — roughly 3.5σ against the ±0.23 error bars — and it is slower on both prefill and decode. It ships for completeness; it is not the "quality" option here. Tier ranking is model-dependent and has to be measured, not assumed.
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
> ### ⚠️ STOCK `llama.cpp` WILL NOT LOAD THIS MODEL > 🚀 **37.86 tok/s** on a 119B MoE — **63.07 GiB**, 7 GiB smaller than UD-Q4_K_XL. A 4-bit **ROCmFP4** quantization for **AMD gfx1151 (Ryzen AI MAX+ 395 / Strix Halo)**, quantized from **BF16 (222 GiB)** — a lossless source, not a requantization of a lower-bit build. `Q4_0_ROCMFP4_COHERENT` (ftype 102) exists only in `charlie12345/ROCmFPX`, not upstream llama.cpp. Ignore the auto-generated "Use this model" commands above.
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
The ROCmFP4 files already in this repository (STRIX, STRIX_LEAN and FAST) are built without importance weighting. This file keeps the STRIX recipe but adds importance-matrix weighting from an agentic calibration imatrix, and raises the output head and the token embedding to Q6_K. The numbers below pair it with the DFlash2 draft head rather than the in-repo MTP head. **How it was built:** - Tool: `gufo-quant`, recipe STRIX, weighted with the agentic imatrix. - Tensor overrides: `output.weight` and `token_embd.weight` at Q6_K; census 866 tensors (82 at type 100, 421 at type 101). - The exact command line was not recorded. **Measured on Super ROC** (AMD Ryzen AI Max+ 395 (Radeon 8060S, gfx1151), 128 GB, ROCm 7.2.4; one request at a time, temperature 0, cold prompt cache):
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
Uncensored ROCmFP4 quantisations of Qwen/Qwen3.8-27B, built with the **same ROCmFPX pipeline and the same card ftypes** as `kingjones777/Qwen3.8-27B-ROCmFP4-STRIX-MTP-GGUF`. **Research artifact.** Abliteration removes content-refusal. It does not add capability. Do not ship this as a product default. The aligned repo remains the serving default. The aligned card's finding still stands: **the lever is MTP depth, not the ftype.** llama.cpp's default `--spec-draft-n-max 16` is about half the achievable throughput on this model. The knee is **n=4**. These uncensored files are the card's FAST (103) and STRIX_LEAN (106) from an abliterated BF16, plus Q6 AGENT/LEAN.
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。
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。