Tev1 (experimental) — Together AI 的本地 AI 模型
作者： Jakub Rusinowski · 最后更新：
Together AI 受 Jev 启发的实验性决策模型，是对 Qwen3.5 的监督微调，从结构化决策中选出一个选项。可在 Ollama 中使用。
变体
Tev1 (experimental) 最小的变体在 Q4_K_M 下约需 1 GB 显存——量化权重加框架开销，不含 KV 缓存。
| 模型 | Q4 下显存 | 显存 | 上下文 | 运行 |
|---|---|---|---|---|
| Tev1 4B → 4B | ~3.6 GB | Not published | ollama pull tev1 | |
| Tev1 0.8B → 0.8B | ~1.3 GB | Not published | ollama pull tev1:0.8b |
显存为 Q4_K_M 下的量化权重加开销，与 GPU 与显存检测器使用同一引擎计算。
如何在本地运行 Tev1 (experimental)
安装 Ollama，然后拉取标签。
ollama pull tev1在上方选择一个尺寸，查看它自己的显存、速度估算和安装命令。
推荐 GPU
目录中能在本地运行 Tev1 (experimental)（至少 1 GB 显存）的最便宜 GPU 是 Intel Arc B570 (10 GB).
Tev1 (experimental) — 常见问题
How much VRAM does Tev1 (experimental) need?
Tev1 (experimental) needs about 1 GB VRAM at Q4_K_M quantization for its smallest variant. Variants: Tev1 4B (4 GB, Q4_K_M); Tev1 0.8B (1 GB, Q4_K_M). On Apple Silicon, unified memory counts toward this requirement.
Can I run Tev1 (experimental) on an RTX 4090 (24 GB)?
Yes — Tev1 (experimental) runs on an RTX 4090 (24 GB) and other 24 GB cards such as the RTX 3090. Smaller variants also fit comfortably on 8–16 GB GPUs at Q4_K_M.
What quantization should I use for Tev1 (experimental)?
Q4_K_M is the best balance of quality and VRAM for Tev1 (experimental) in most cases. Choose Q8_0 for near-lossless quality if you have spare VRAM, or smaller quants (Q3/Q2) only when memory is tight.
How do I run Tev1 (experimental) locally?
Tev1 (experimental) is a decision model: it is called over an HTTP endpoint (/v1/systemone), not chatted with. Where a variant is available in Ollama 0.35 or newer, pull it with `ollama pull` and send requests to the local server; the others ship their own server. Each variant page shows the exact commands for that model.