vllm-sr/d3 — new model trending #24 on Hugging Face
vLLM Semantic Router released d3, a 27B multimodal model that returns calibrated decisions instead of generated text.
vllm-sr/d3 is the 26.09B Apache-2.0 multimodal decision model for vLLM Semantic Router, including a 0.46B vision encoder. It takes text or JSON plus images and videos and returns probabilities for choice, yes/no, and score questions in one call without generating text. On the official Jev Decision Index 0.3.2 text board it scores 66.3, first place and 10.4 above Decision 2.0; the public suite is 65.45. Internal Perception Test multiple-choice video QA is 77.4, and median text latency is 54.4 ms on one AMD Instinct MI325X.
- 26.09B Apache-2.0 model, including a 0.46B vision encoder.
- Scores 66.3 on Jev Decision Index 0.3.2, up 10.4 from Decision 2.0.
- Answers choice, yes/no, and score questions with probabilities, no generated text.
- Perception Test video QA score is 77.4; text median latency is 54.4 ms on one MI325X.
Full article2,964 words · extracted from huggingface.co · click to collapse
# d3
**d3** is the 27B multimodal foundation decision model of Decision 3.0, the decision models of [vLLM Semantic Router](https://github.com/vllm-project/semantic-router). Give it an input (text or JSON, optionally with images and videos) and the questions you need answered: pick one of several options, say yes or no, or rate on a scale. It answers them all in one call and returns a probability for every answer, without generating text.
**Try it free** at [decision.vllm-sr.ai](https://decision.vllm-sr.ai) · Official Space: [vllm-sr/Decision-3.0](https://huggingface.co/spaces/vllm-sr/Decision-3.0)
| | |
| --- | --- |
| **Parameters** | 26.09B, including the 0.46B vision encoder |
| **Inputs** | Text or JSON, plus images and videos (several per request) |
| **Decision types** | Choice · Yes / No · Score |
| **License** | Apache-2.0 |
## Highlights
- **Jev Decision Index 0.3.2: 66.3, first on the official text board** and on its Pareto frontier; public suite 65.45, measured with the official 0.3 kit on the released weights: all 140,178 public requests answered, none unsupported.
- **+10.4 over Decision 2.0** (its 27B model: 55.9 on the official board): +8.5 on the public suite, ahead in all five of its areas, and +7.1 and +12.6 on the private same-skill tests and new-domain tasks.
- **Reads images:** multiple images per request (PNG, JPEG or WebP), given as paths, URLs, PIL images or base64 data URLs; every question of the request sees all of them.
- **Reads videos:** multiple videos per request (MP4, WebM, MOV or MKV), given as paths, URLs, base64 data URLs or frame arrays, read at 2 frames per second; **Perception Test (multiple-choice video QA): 77.4** (internal evaluation).
- **Speed:** text requests take a median of 54.4 ms, requests with an image a median of 290 ms, and requests with a 10-second video a median of 500 ms, on one AMD Instinct MI325X, one request at a time with AutoModel.
- **Serving:** 27 requests/s with p99 under 1 s on one AMD Instinct MI325X with vLLM Semantic Router, 1.5× AutoModel (one-question text, with batching, now the default); with one request per forward, the former default, it gave AutoModel's answers byte for byte when both used the same kernel settings ([Serving performance](#serving-performance)).
- **Many questions, one call:** Choice, Yes / No and Score questions about the same input are answered together, each from its own forward pass over the input, with a probability for every option.
## Quickstart
```bash