AI NEWS.ccAI正在发生 未来正在更新
♡ 我的收藏 0
AI NEWS · 模型情报站发现能力边界，让每一次选择更有依据 ↗基础档案先收录 · 缺失资料持续补齐
THE AI MODEL DIRECTORY 模型探索，从这里开始

找到适合你的 AI 模型。

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。

热门探索
84,996 款模型档案13202 家厂商 / 发布者45 个能力分类465 款有可比报价 · 30 款有公开评测
T文本生成▧图像生成▷视频生成⌘代码生成◖语音合成♫音乐生成⬡3D 建模▤文档分析
DISCOVER & COMPARE

探索 AI 模型 255

为你发现值得关注的模型

TADA-3B-ML is a multilingual text-to-speech model built on Meta Llama 3.2 3B with a flow-matching (FM) speech decoder and custom Hume codec. Key property: **1:1 token alignment** — every text token maps to exactly one speech vector, eliminating transcript hallucination. Supports 10 languages (en, es, ja, zh, de, fr, it, pt, ko, ar). 24 kHz mono output. The Q4_K file uses a TADA-aware quantization policy (tail=14): large transformer block matrices are quantized, while the last 14 token-embedding rows and all `tada.*` flow-matching tensors are kept at F16. This preserves the timing and acoustic conditioning paths where quantization noise matters most. For near-lossless quality use `tada-tts-3b-ml-q8_0.gguf`; for the F16 reference quality use `tada-tts-3b-ml-f16.gguf`.

译文排队中 · 当前显示来源原文

语音合成
输入 文本→ 输出 音频
价格待核验多模态与渠道费用须分别核对
查看详情 ↗
来源 / 文档 ↗

Apache-2.0 Tesseract LSTM line-recognition models converted to GGUF for CrispEmbed. The repository contains English, German, French, Spanish, Italian, Portuguese, Dutch, Russian, Arabic, Simplified Chinese, Japanese, and Korean variants. These are line recognizers: normalize and crop individual text lines before recognition, or pair them with a detector such as DBNet for full-page OCR. Each language has an F32 reference, an F16 deployment model, and Q8_0/Q4_K variants. The quantized models preserve `output.weight` and `output.bias` at the source precision (F16 for the multilingual F16-derived models; F32 for German regenerated from its original `.traineddata`). Only recurrent matrices are quantized. This preserves the CTC decision boundary and keeps critical character logits stable.

译文排队中 · 当前显示来源原文

图像描述
输入 图像→ 输出 文本
价格待核验多模态与渠道费用须分别核对
查看详情 ↗
来源 / 文档 ↗

> The Q8_0/Q4_K files keep the diffusion prediction head, connectors and > EOS classifier at full precision (only the two Qwen2 backbones are > quantized). The head runs under classifier-free guidance over 20 solver > steps, so quantizing it could push the first frames onto a wrong > trajectory that decodes as a brief non-speech "music"/hum onset before > the voice; keeping it at full precision avoids that at a small size cost. A voice prompt is **required** for TTS. Each `.gguf` voice pack ships pre-computed KV caches that establish a fixed speaker identity (the realtime variant is *not* WAV-cloning — for runtime cloning, use `cstr/vibevoice-1.5b-GGUF` instead). This repo bundles all 25 demo voices from `microsoft/VibeVoice@main demo/voices/streaming_model/` — same MIT license as the model.

译文排队中 · 当前显示来源原文

语音合成
输入 文本→ 输出 音频
价格待核验多模态与渠道费用须分别核对
查看详情 ↗
来源 / 文档 ↗

Historical T4 Vulkan component measurements found CFM per audio step of 70.2 ms for Q8_0 versus 59.4 ms for the mixed recipe at ten steps. An eight-step run measured 46.7 ms. These component timings do **not** establish speech quality or an accepted end-to-end speedup. The mixed file was built with `crispasr-quantize voxcpm2-f16.gguf out.gguf q8_0 --tensor-type "^locdit\.=f16"`. A previous CPU synthesis roundtrip is not GPU acceptance. **Validation update, 2026-10-09:** matched real T4 tests at ten steps and seed 2 produced repeated tokens in the **Nemotron ASR transcripts** of the short test sentence for both Q8_0 and the mixed recipe. Six of sixteen exact primary ASR cases passed. This rejected gate is not proof that synthesized audio audibly stutters: a later fixed-audio comparison found **both Parakeet and Qwen3 transcribe every one of 22 original Python/native CPU/GPU controls exactly** across three resampling paths, while Nemotron retains errors. Resampling alone does not resolve the discrepancy. The completed original NVIDIA float32 and native F16 comparison reads **all 22 VoxCPM2 controls correctly**; native Q4 reads only 8/22. Original and F16 normalized words agree on all 26 tested clips, and fresh/reused sessions agree. The repeated tokens are therefore attributable to the native Nemotron Q4 recognizer on these fixed controls. Critical RNNT precision guards are being tested before changing that quant or its defaults. The published Q4 primary gate remains rejected while that recognizer is repaired. This diagnosis does not certify eight steps, Intel B390 performance, or complete original/native waveform parity. Defaults remain ten steps. Original NVIDIA/F16/Q4 inputs, features, encoder states and complete transcripts (SHA256 `7762f2a73d8ae7dad05d667fbf0bca393a0b3b7ad622b1b1f09473fa8f5d0463`).

译文排队中 · 当前显示来源原文

语音合成
输入 文本→ 输出 音频
价格待核验多模态与渠道费用须分别核对
查看详情 ↗
来源 / 文档 ↗

GGUF quantizations of mistralai/Voxtral-Mini-4B-Realtime-2602, a **4.4B-parameter realtime streaming speech-to-text model** with a causal audio encoder and configurable transcription delay. ⚠ This table used to name the F16 `voxtral-mini-4b-realtime.gguf`, and no such file was ever uploaded. The published name is `…-f16.gguf`, matching the quant suffix CrispASR's `-m auto:f16` resolver expects.

译文排队中 · 当前显示来源原文

语音识别
输入 音频→ 输出 文本
价格待核验多模态与渠道费用须分别核对
查看详情 ↗
来源 / 文档 ↗
COST EXPLORER

好模型，也要算好成本。

费用分析

调整月度 Token 用量，直观比较不同模型的预估 API 开支。

1 M = 100 万 Token · 美元计价
仅比较同一渠道的基础文本 Token 报价

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。

MODEL HANDBOOK

读懂模型，再做选择

选型与常见问题
API 价格应该怎么比较？+

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。

开放权重，就代表可以免费商用吗？+

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。

排行榜第一，就是最适合我的模型吗？+

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。

AI NEWS · 从信息到选择

资讯看趋势，工具提效率，模型筑可能。

探索 AI 资讯 ↗发现 AI 工具 ↗