Cerebras 上线 Qwen3.8-27B 及 GPT-OSS 推理 API
Cerebras 上线 Qwen3.8-27B 模型推理服务
Cerebras 接入主流开源模型并提供高吞吐 API，对需要低成本、低延迟推理服务的开发者是重要基础设施更新，建议关注其性能表现。
Copy pageCopy MCP ServerView as Markdown
复制页面复制 MCP ServerView as Markdown
Models on Cerebras public endpoints are available on the free trial and pay-as-you-go tiers, subject to rate limits and pricing. For additional model families, reserved capacity, higher throughput, and production SLAs, see Dedicated Endpoints.
Cerebras 公共端点上的模型在免费试用和按量付费层级中可用，受速率限制和定价约束。如需更多模型系列、预留容量、更高吞吐量以及生产环境 SLA，请参阅 Dedicated Endpoints。
New here? Follow the Quickstart to make your first API call. To pick a model by use case, see the model selection guide. Select any model name below for full specs, capabilities, and per-tier limits.
新用户？请按照快速入门指南进行首次 API 调用。如需按用例选择模型，请参阅模型选择指南。点击下方任意模型名称可查看完整规格、功能及各级别限制。
Available Models
可用模型
| Model Name | Model ID | Parameters | Context (free / paid) | Speed (tokens/s) |
|---|---|---|---|---|
| OpenAI GPT OSS | gpt-oss-120b | 120 billion | 65k / 131k | ~3000 |
| Qwen 3.8 27B | qwen-3.8-27b | 27 billion | 64k / 128k | ~1500 |
| 模型名称 | 模型 ID | 参数量 | 上下文长度（免费/付费） | 速度（tokens/s） |
|---|---|---|---|---|
| OpenAI GPT OSS | gpt-oss-120b | 1200亿 | 65k / 131k | ~3000 |
| Qwen 3.8 27B | qwen-3.8-27b | 270亿 | 64k / 128k | ~1500 |
Looking for more models? Many additional model families are available through Dedicated Endpoints.
寻找更多模型？通过 Dedicated Endpoints 可使用许多其他模型系列。
Model Compression
模型压缩
This section provides transparency about the compression state of each model available on our platform. We host a variety of open-source models from the community. We do not currently host pruned models on our public endpoints. All models served through our public endpoints are the original, unpruned versions. While we conduct research on pruning techniques like REAP (Router-weighted Expert Activation Pruning), these pruned models are shared with the research community on Hugging Face but are not available through our shared API. You can read more about REAP in our research blog. All of our public models are unpruned. Cerebras uses selective weight-only quantization only during storage to preserve maximal quality. This means that the weights are stored in partial 16-bit / 8-bit / 4-bit, in-line with industry standards. For quality, sensitive layers are stored at full precision with dequantization on the fly, so operations are done in high precision. The activations, attention, and kv cache remain in full precision and unquantized.
本节旨在透明地展示我们平台上各模型的压缩状态。我们托管了来自社区的多种开源模型。目前，我们的公共端点不托管剪枝模型。所有通过公共端点提供的模型均为原始未剪枝版本。虽然我们正在进行如 REAP（基于路由器权重的专家激活剪枝）等剪枝技术研究，但这些剪枝模型仅在 Hugging Face 上与研究社区共享，无法通过我们的共享 API 获取。您可以在我们的研究博客中了解更多关于 REAP 的信息。我们所有的公开模型均未剪枝。Cerebras 仅在存储阶段使用选择性仅权重量化以保留最大质量。这意味着权重以部分 16 位/8 位/4 位格式存储，符合行业标准。为保证质量，敏感层以全精度存储并实时反量化，因此运算在高精度下进行。激活值、注意力机制和 kv cache 保持全精度且未量化。
Frequently Asked Questions
常见问题解答
Will you change a model's architecture without notice?
您是否会未经通知更改模型架构？
No. We are committed to serving the original models for all existing endpoints, without modification. We do not alter model architectures via pruning on our hosted portfolio. If we explore additional compression techniques (like pruning) in the future, these would be offered as separate endpoints with pruning-specific names, ensuring complete transparency and allowing you to choose which version best fits your needs.
不会。我们承诺为所有现有端点提供原始模型服务，不作任何修改。我们不会通过剪枝来改变托管组合中的模型架构。如果未来我们探索其他压缩技术（如剪枝），这些将以带有特定剪枝名称的独立端点形式提供，确保完全透明，并让您可以选择最适合需求的版本。
Where can I find your REAP pruned models?
在哪里可以找到您的 REAP 剪枝模型？
Our REAP pruned models are available on Hugging Face for research and experimentation purposes: Cerebras REAP Collection. These models demonstrate our pruning research but are not served through our production API.
我们的 REAP 剪枝模型已在 Hugging Face 上提供，供研究和实验使用：Cerebras REAP Collection。这些模型展示了我们的剪枝研究成果，但不会通过我们的生产 API 提供服务。
What are compression, quantization, and pruning?
什么是压缩、量化和剪枝？
Compression is an umbrella term for techniques that reduce model size or computational requirements. Common compression techniques include:
压缩是一个涵盖性术语，指用于减小模型规模或降低计算需求的技术。常见的压缩技术包括：
- Quantization: Reducing the precision of numbers used to represent model weights (e.g., converting from FP16 to FP8). This reduces memory usage without changing the model’s architecture.
- Pruning: Permanently removing parts of a model, like layers or experts, to reduce model size. This changes the model’s architecture and creates a different model.
- 量化：降低用于表示模型权重的数字精度（例如，从 FP16 转换为 FP8）。这可以在不改变模型架构的情况下减少内存占用。
- 剪枝：永久移除模型的某些部分（如层或专家），以减小模型规模。这会改变模型的架构并生成一个不同的模型。
Was this page helpful?
此页是否有帮助？
YesNo
是 否
⌘I
⌘I
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力