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。
*👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord*. **Model creator:** google **Original model**: gemma-3n-E4B-it **GGUF quantization:** provided by LM Studio team using `llama.cpp` 🙏 Special thanks to Georgi Gerganov and the whole team working on llama.cpp for making all of this possible.
译文排队中 · 当前显示来源原文
*👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord*. **Model creator**: google **Original model**: gemma-4-12B-it **GGUF quantization**: provided by LM Studio team using `llama.cpp` release b10069 🙏 Special thanks to Georgi Gerganov and the whole team working on llama.cpp for making all of this possible.
译文排队中 · 当前显示来源原文
_👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord_. **Model creator**: google **Original model**: gemma-4-12B-it **MLX quantization**: provided by LM Studio team using mlx_vlm 4-bit quantized version of gemma-4-12B-it using MLX, optimized for Apple Silicon.
译文排队中 · 当前显示来源原文
_👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord_. **Model creator**: google **Original model**: gemma-4-12B-it **MLX quantization**: provided by LM Studio team using mlx_vlm 5-bit quantized version of gemma-4-12B-it using MLX, optimized for Apple Silicon.
译文排队中 · 当前显示来源原文
_👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord_. **Model creator**: google **Original model**: gemma-4-12B-it **MLX quantization**: provided by LM Studio team using mlx_vlm 6-bit quantized version of gemma-4-12B-it using MLX, optimized for Apple Silicon.
译文排队中 · 当前显示来源原文
_👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord_. **Model creator**: google **Original model**: gemma-4-12B-it **MLX quantization**: provided by LM Studio team using mlx_vlm 8-bit quantized version of gemma-4-12B-it using MLX, optimized for Apple Silicon.
译文排队中 · 当前显示来源原文
*👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord*. **Model creator**: google **Original model**: gemma-4-12B-it-qat-q4_0-unquantized **GGUF quantization**: provided by LM Studio team using `llama.cpp` release b10069 🙏 Special thanks to Georgi Gerganov and the whole team working on llama.cpp for making all of this possible.
译文排队中 · 当前显示来源原文
*👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord*. **Model creator**: google **Original model**: gemma-4-26B-A4B-it **GGUF quantization**: provided by LM Studio team using `llama.cpp` release b10069 🙏 Special thanks to Georgi Gerganov and the whole team working on llama.cpp for making all of this possible.
译文排队中 · 当前显示来源原文
_👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord_. **Model creator**: google **Original model**: gemma-4-26B-A4B-it **MLX quantization**: provided by LM Studio team using mlx_vlm 4-bit quantized version of gemma-4-26B-A4B-it using MLX, optimized for Apple Silicon.
译文排队中 · 当前显示来源原文
_👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord_. **Model creator**: google **Original model**: gemma-4-26B-A4B-it **MLX quantization**: provided by LM Studio team using mlx_vlm 5-bit quantized version of gemma-4-26B-A4B-it using MLX, optimized for Apple Silicon.
译文排队中 · 当前显示来源原文
_👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord_. **Model creator**: google **Original model**: gemma-4-26B-A4B-it **MLX quantization**: provided by LM Studio team using mlx_vlm 6-bit quantized version of gemma-4-26B-A4B-it using MLX, optimized for Apple Silicon.
译文排队中 · 当前显示来源原文
_👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord_. **Model creator**: google **Original model**: gemma-4-26B-A4B-it **MLX quantization**: provided by LM Studio team using mlx_vlm 8-bit quantized version of gemma-4-26B-A4B-it using MLX, optimized for Apple Silicon.
译文排队中 · 当前显示来源原文
*👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord*. **Model creator**: google **Original model**: gemma-4-26B-A4B-it-qat-q4_0-unquantized **GGUF quantization**: provided by LM Studio team using `llama.cpp` release b10069 🙏 Special thanks to Georgi Gerganov and the whole team working on llama.cpp for making all of this possible.
译文排队中 · 当前显示来源原文
_👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord_. **Model creator**: google **Original model**: gemma-4-26B-A4B-it-QAT **MLX quantization**: provided by LM Studio team using mlx_vlm 4-bit quantized version of gemma-4-26B-A4B-it-QAT using MLX, optimized for Apple Silicon.
译文排队中 · 当前显示来源原文
*👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord*. **Model creator**: google **Original model**: gemma-4-31B-it **GGUF quantization**: provided by LM Studio team using `llama.cpp` release b10069 🙏 Special thanks to Georgi Gerganov and the whole team working on llama.cpp for making all of this possible.
译文排队中 · 当前显示来源原文
_👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord_. **Model creator**: google **Original model**: gemma-4-31B-it **MLX quantization**: provided by LM Studio team using mlx_vlm 4-bit quantized version of gemma-4-31B-it using MLX, optimized for Apple Silicon.
译文排队中 · 当前显示来源原文
_👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord_. **Model creator**: google **Original model**: gemma-4-31B-it **MLX quantization**: provided by LM Studio team using mlx_vlm 5-bit quantized version of gemma-4-31B-it using MLX, optimized for Apple Silicon.
译文排队中 · 当前显示来源原文
_👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord_. **Model creator**: google **Original model**: gemma-4-31B-it **MLX quantization**: provided by LM Studio team using mlx_vlm 6-bit quantized version of gemma-4-31B-it using MLX, optimized for Apple Silicon.
译文排队中 · 当前显示来源原文
_👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord_. **Model creator**: google **Original model**: gemma-4-31B-it **MLX quantization**: provided by LM Studio team using mlx_vlm 8-bit quantized version of gemma-4-31B-it using MLX, optimized for Apple Silicon.
译文排队中 · 当前显示来源原文
*👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord*. **Model creator**: google **Original model**: gemma-4-31B-it-qat-q4_0-unquantized **GGUF quantization**: provided by LM Studio team using `llama.cpp` release b10069 🙏 Special thanks to Georgi Gerganov and the whole team working on llama.cpp for making all of this possible.
译文排队中 · 当前显示来源原文
*👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord*. **Model creator**: google **Original model**: gemma-4-E2B-it **GGUF quantization**: provided by LM Studio team using `llama.cpp` release b10076 🙏 Special thanks to Georgi Gerganov and the whole team working on llama.cpp for making all of this possible.
译文排队中 · 当前显示来源原文
_👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord_. **Model creator**: google **Original model**: gemma-4-E2B-it **MLX quantization**: provided by LM Studio team using mlx_vlm 4-bit quantized version of gemma-4-E2B-it using MLX, optimized for Apple Silicon.
译文排队中 · 当前显示来源原文
_👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord_. **Model creator**: google **Original model**: gemma-4-E2B-it **MLX quantization**: provided by LM Studio team using mlx_vlm 5-bit quantized version of gemma-4-E2B-it using MLX, optimized for Apple Silicon.
译文排队中 · 当前显示来源原文
_👾 LM Studio Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on Discord_. **Model creator**: google **Original model**: gemma-4-E2B-it **MLX quantization**: provided by LM Studio team using mlx_vlm 6-bit quantized version of gemma-4-E2B-it using MLX, optimized for Apple Silicon.
译文排队中 · 当前显示来源原文
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。
