# Sage S2.1 Flash

> Sage S2.1 Flash is a scientific multimodal model from Shanghai Artificial Intelligence Laboratory for iterative research. It combines scientific data understanding with retrieval, coding and tool use, takes text and images, and has a 512K-token context window.

**ID**: `intern/sage-s2.1-flash`  
**Creator**: intern  
**Category**: multimodal  
**Context**: 524K tokens  
**Pricing**: BYOK — billed by your provider  
**Released**: 2026-11-16  
**Links**: [Weights on Hugging Face](https://huggingface.co/internlm/Sage-S2.1-Flash) · [Announcement](https://x.com/AtriaASI/status/2122319410623029578) · [AnyRouter release](https://anyrouter.dev/blog/releases/2026-11-17-sage-s2-1-flash)  
**Web page**: https://anyrouter.dev/model/intern/sage-s2.1-flash

**Input modalities**: text, image  
**Output modalities**: text  

**Capabilities**: chat, streaming, vision, reasoning, function-calling, coding, agentic

## Usage

**Endpoint**: `POST https://anyrouter.dev/api/v1/chat/completions`

```bash
curl https://anyrouter.dev/api/v1/chat/completions \
  -H "Authorization: Bearer $ANYROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "intern/sage-s2.1-flash",
  "messages": [
    {
      "role": "user",
      "content": "Say hi in 3 words."
    }
  ]
}'
```

## Providers

| Provider | Input | Output |
| --- | --- | --- |
| intern-byok | BYOK — billed by your provider | — |

BYOK routes are billed by your own provider key — AnyRouter charges $0. Rates marked (list) are the provider's published list price.

## Model card

- Native context: 524288 tokens [2]
- License: MIT, open weights (https://huggingface.co/internlm/Sage-S2.1-Flash) [1]

### Strengths [1,3]

- Built for iterative scientific research with retrieval, coding and tool use.
- Reads more than ten scientific input types through Sage SciPerception, including molecular graphs, crystal structures and protein structures.

### Known limits [1,2]

- The hosted Intern API accepts text and image input; other scientific input types require the open-weight projectors.
- Vendor reports a limited-time discount on TokenPlan rates.

### Use cases [1,3]

- Scientific literature and data analysis
- Materials and molecular property prediction
- Research agents and document workflows

### API notes [2,1]

- Hosted on the Intern Discovery API at POST /v1/chat/completions with a Bearer TokenPlan key.
- The host also exposes OpenAI Responses and Anthropic Messages protocols.
- Context is 512K tokens; input types on the hosted API are text and images.

### Links

- announcement: https://x.com/AtriaASI/status/2122319410623029578
- model card: https://huggingface.co/internlm/Sage-S2.1-Flash
- docs: https://discovery.intern-ai.org.cn/token-plan/home

### Sources

1. [Sage S2.1 Flash model card](https://huggingface.co/internlm/Sage-S2.1-Flash/raw/main/README.md) (accessed 2026-11-17)
2. [Intern TokenPlan models and pricing](https://discovery.intern-ai.org.cn/token-plan/home) (accessed 2026-11-17)
3. [Atria ASI announcement of Sage S2.1 Flash](https://x.com/AtriaASI/status/2122319410623029578) (accessed 2026-11-17)

## Source

- Upstream docs: https://huggingface.co/internlm/Sage-S2.1-Flash
