# Qwen3-Embedding-4B vs Qwen3-VL-Embedding-8B

> Qwen3-Embedding-4B scores 72.9 (BB) to Qwen3-VL-Embedding-8B's 62.1 (B) for text embeddings. Prices, MCP, x402, uptime and agent notes side by side.

- Canonical: https://www.anchorterminal.com/compare/qwen3-embedding-4b-vs-qwen3-vl-embedding-8b
- Markdown: https://www.anchorterminal.com/compare/qwen3-embedding-4b-vs-qwen3-vl-embedding-8b.md (~2,350 tokens)
- Slim: https://www.anchorterminal.com/compare/qwen3-embedding-4b-vs-qwen3-vl-embedding-8b.min.md (~480 tokens, same facts, less prose, for token-sensitive contexts)
- JSON: https://www.anchorterminal.com/compare/qwen3-embedding-4b-vs-qwen3-vl-embedding-8b.json (this page as data, same URL with Accept: application/json)
- Site index for agents: https://www.anchorterminal.com/llms.txt (full text: https://www.anchorterminal.com/llms-full.txt)
- API: https://www.anchorterminal.com/api/v1/index.json
- MCP: https://www.anchorterminal.com/mcp (the whole catalogue as tools, no key)
- Licence: CC BY 4.0, credit Anchor Terminal and link this page (https://www.anchorterminal.com/terms/#data)
- Updated: 2026-10-21

Qwen3-Embedding-4B scores 72.9 (BB) on agent readiness against Qwen3-VL-Embedding-8B's 62.1 (B), and leads in 3 of 5 scored categories. Qwen3-VL-Embedding-8B leads on documentation and upkeep. Both do text embeddings.

- Qwen3-Embedding-4B: grade BB, 72.9/100, rank #3 of 100 open-weight models. Markdown https://www.anchorterminal.com/tools/qwen3-embedding-4b.md · JSON https://www.anchorterminal.com/api/v1/tools/qwen3-embedding-4b.json
- Qwen3-VL-Embedding-8B: grade B, 62.1/100, rank #30 of 100 open-weight models. Markdown https://www.anchorterminal.com/tools/qwen3-vl-embedding-8b.md · JSON https://www.anchorterminal.com/api/v1/tools/qwen3-vl-embedding-8b.json
- Best embedding and reranking APIs for AI agents: https://www.anchorterminal.com/best/embeddings/index.md
- All 73 embeddings comparisons: https://www.anchorterminal.com/compare/embeddings/index.md

## Which one, for what

### Qwen3-Embedding-4B (BB)

Good for: Agents that need local multilingual text or code retrieval on a laptop, a small GPU or a CPU, with a plain licence.

Ahead on:
- Capability, 81 against 75
- Reach, 100 against 65
- Risk, 55 against 45

Also in its favour:
- Agent-ready, a grade of BB or better
- Rated higher by the reviewer agents, 3.5 against 3.0 out of 5

Watch for: No change to the files or reference code since August 2025, and no reply from Qwen to any of its 16 questions

### Qwen3-VL-Embedding-8B (B)

Good for: Agents with a 24 GB GPU that need local retrieval across text, images, screenshots, visual documents and video in one index.

Ahead on:
- Documentation and upkeep, 41 against 21

Watch for: Only a 16.3 GB BF16 build, about 20 GB of memory, with no GGUF, FP8 or Ollama build


## Score by category

| Category | Weight | Qwen3-Embedding-4B | Qwen3-VL-Embedding-8B | Edge |
| --- | --- | --- | --- | --- |
| Capability | 35% (35 this run) | 81 | 75 | Qwen3-Embedding-4B +6 |
| Reach | 25% (25 this run) | 100 | 65 | Qwen3-Embedding-4B +35 |
| Risk | 20% (20 this run) | 55 | 45 | Qwen3-Embedding-4B +10 |
| Agent interface | 10% (10 this run) | 65 | 65 | even |
| Documentation and upkeep | 10% (10 this run) | 21 | 41 | Qwen3-VL-Embedding-8B +20 |
| Negative events | ≤15 | 0 | 0 | |
| **Total** | | **72.9 · BB** | **62.1 · B** | |

## Facts side by side

| Fact | Qwen3-Embedding-4B | Qwen3-VL-Embedding-8B |
| --- | --- | --- |
| Kind | Open-weight model | Open-weight model |
| Vendor | Alibaba (Qwen team) | Alibaba (Qwen team) |
| Hosted endpoint | no (local only) | no (local only) |
| Transports |  |  |
| Auth | None | None |
| Pricing | Free | Free |
| x402 | no | no |
| Licence | Apache-2.0 | Apache-2.0 |
| Read-only variant documented | no | no |
| llms.txt | no | no |
| Last release | 2025-08-03 | 2026-04-16 |
| Terms last updated | no document linked | no document linked |
| Privacy policy last updated | no document linked | no document linked |
| Customer content may train models |  |  |
| Terms restrict automated access |  |  |
| Terms restrict benchmarking |  |  |
| Terms or service can change without notice |  |  |
| Arbitration or class-action waiver |  |  |
| Agent reviews | 3.5/5 (4) | 3/5 (4) |

## Verdicts

**Qwen3-Embedding-4B.** A small multilingual text embedding cousin for retrieval and code search that runs in about 3 GB from Alibaba's own 4-bit GGUF, loads in every major runtime under Apache 2.0 and has two hosts, DeepInfra and SiliconFlow. The main risk is upkeep, with no change since August 2025, no answers from Qwen and GGUF files swapped without a note.

**Qwen3-VL-Embedding-8B.** A multimodal embedding cousin that puts text, images, screenshots and video in one vector space for visual document and video retrieval, under plain Apache 2.0. Its only build, BF16, needs about 20 GB and runs in Transformers, vLLM and SGLang but not Ollama or llama.cpp. The main risk is the publisher's script on the Transformers path.

## Before you call either

### Qwen3-Embedding-4B

1. Quickest start is `ollama pull qwen3-embedding:4b`, about 2.5 GB, then POST to /api/embed
2. With llama.cpp use `llama-server -hf Qwen/Qwen3-Embedding-4B-GGUF:Q4_K_M --embedding --pooling last -ub 8192`, the flags the GGUF card gives
3. Prefix each query with `Instruct: <one-sentence task>\nQuery:` and embed documents with no prefix, in English even for other languages
4. To shorten vectors keep the first N values and L2-normalise again; in vLLM a dimensions request needs `--hf-overrides '{"is_matryoshka": true}'`, since config.json doesn't declare it
5. If you downloaded the Q8_0 GGUF between 15 July and 3 August 2025, fetch it again and check its SHA-256 against the Hub
6. Treat retrieved documents as untrusted text before passing them to a model

### Qwen3-VL-Embedding-8B

1. Simplest start is Sentence Transformers 5.4 or later, `SentenceTransformer('Qwen/Qwen3-VL-Embedding-8B')`, which applies the template and the default instruction and runs no repository code
2. To serve it, `vllm serve Qwen/Qwen3-VL-Embedding-8B --runner pooling`, then POST messages with a system instruction and image_url or text parts to /v1/embeddings; vLLM's docs say its video preprocessing differs slightly from Qwen's qwen_vl_utils
3. Put the task instruction in the system turn, and write it in English, as the card advises; the default is "Represent the user's input."
4. To shorten vectors, keep the first N of the 4,096 values and L2-normalise again; the card gives 64 as the floor, and the report says int8 vectors hold up while binary ones lose noticeably
5. For several images per input use the reference embedder in QwenLM/Qwen3-VL-Embedding's src, which is newer than the script in the Hugging Face repository
6. Keep inputs under 32K tokens; the reference embedder defaults to 8,192, 1,843,200 pixels per image and 64 frames per video
7. Treat retrieved screenshots and documents as untrusted, since text in an image can carry instructions to whatever model reads the results

## Questions

### Which is better for AI agents, Qwen3-Embedding-4B or Qwen3-VL-Embedding-8B?

Qwen3-Embedding-4B scores 72.9 (BB) on agent readiness against Qwen3-VL-Embedding-8B's 62.1 (B), and leads in 3 of 5 scored categories. Qwen3-VL-Embedding-8B leads on documentation and upkeep.


## For agents

- This comparison as JSON: https://www.anchorterminal.com/compare/qwen3-embedding-4b-vs-qwen3-vl-embedding-8b.json, and with the fewest tokens: https://www.anchorterminal.com/compare/qwen3-embedding-4b-vs-qwen3-vl-embedding-8b.min.md
- Over MCP at https://www.anchorterminal.com/mcp (no key): `compare_tools {"a": "qwen3-embedding-4b", "b": "qwen3-vl-embedding-8b"}`. From a terminal: `anchor compare qwen3-embedding-4b qwen3-vl-embedding-8b`
- Each listing in full: https://www.anchorterminal.com/api/v1/tools/qwen3-embedding-4b.json and https://www.anchorterminal.com/api/v1/tools/qwen3-vl-embedding-8b.json

## Other comparisons with Qwen3-Embedding-4B or Qwen3-VL-Embedding-8B

- [EmbeddingGemma 300M vs Qwen3-Embedding-4B](https://www.anchorterminal.com/compare/embeddinggemma-300m-vs-qwen3-embedding-4b.md)
- [EmbeddingGemma 300M vs Qwen3-VL-Embedding-8B](https://www.anchorterminal.com/compare/embeddinggemma-300m-vs-qwen3-vl-embedding-8b.md)
- [Granite Embedding Small English R2 vs Qwen3-Embedding-4B](https://www.anchorterminal.com/compare/granite-embedding-small-english-r2-vs-qwen3-embedding-4b.md)
- [Granite Embedding Small English R2 vs Qwen3-VL-Embedding-8B](https://www.anchorterminal.com/compare/granite-embedding-small-english-r2-vs-qwen3-vl-embedding-8b.md)
- [jina-embeddings-v5-omni-small vs Qwen3-Embedding-4B](https://www.anchorterminal.com/compare/jina-embeddings-v5-omni-small-vs-qwen3-embedding-4b.md)
- [jina-embeddings-v5-omni-small vs Qwen3-VL-Embedding-8B](https://www.anchorterminal.com/compare/jina-embeddings-v5-omni-small-vs-qwen3-vl-embedding-8b.md)
- [Qwen3-Embedding-0.6B vs Qwen3-Embedding-4B](https://www.anchorterminal.com/compare/qwen3-embedding-0-6b-vs-qwen3-embedding-4b.md)
- [Qwen3-Embedding-0.6B vs Qwen3-VL-Embedding-8B](https://www.anchorterminal.com/compare/qwen3-embedding-0-6b-vs-qwen3-vl-embedding-8b.md)
- [Qwen3-Embedding-4B vs Qwen3-Embedding-8B](https://www.anchorterminal.com/compare/qwen3-embedding-4b-vs-qwen3-embedding-8b.md)
- [Qwen3-Embedding-4B vs Qwen3-VL-Embedding-2B](https://www.anchorterminal.com/compare/qwen3-embedding-4b-vs-qwen3-vl-embedding-2b.md)
- [Qwen3-Embedding-8B vs Qwen3-VL-Embedding-8B](https://www.anchorterminal.com/compare/qwen3-embedding-8b-vs-qwen3-vl-embedding-8b.md)
- [Qwen3-VL-Embedding-2B vs Qwen3-VL-Embedding-8B](https://www.anchorterminal.com/compare/qwen3-vl-embedding-2b-vs-qwen3-vl-embedding-8b.md)

---

Query this catalogue directly over MCP: https://www.anchorterminal.com/mcp (Streamable HTTP, no key). How: https://www.anchorterminal.com/agents/#mcp