[BUG] New Qwen embedding models missing for code indexing on openrouter

[dmarkey]

Problem (one or two sentences)

These models have been released on Openrouter and are missing for code indexing. Ensure to research the correct vector size.

Qwen: Qwen3 Embedding 0.6B

103 tokens
The Qwen3 Embedding model series is the latest proprietary model of the Qwen family, specifically designed for text embedding and ranking tasks. This series inherits the exceptional multilingual capabilities, long-text understanding, and reasoning skills of its foundational model. The Qwen3 Embedding series represents significant advancements in multiple text embedding and ranking tasks, including text retrieval, code retrieval, text classification, text clustering, and bitext mining.

by qwen
33K context
$0.01/M input tokens
$0/M output tokens
Qwen: Qwen3 Embedding 4B

24 tokens
The Qwen3 Embedding model series is the latest proprietary model of the Qwen family, specifically designed for text embedding and ranking tasks. This series inherits the exceptional multilingual capabilities, long-text understanding, and reasoning skills of its foundational model. The Qwen3 Embedding series represents significant advancements in multiple text embedding and ranking tasks, including text retrieval, code retrieval, text classification, text clustering, and bitext mining.

by qwen
33K context
$0.02/M input tokens
$0/M output tokens

Context (who is affected and when)

Openrouter users that want to use code embedding with qwen 0.6b and 4b

Reproduction steps

Set up code indexing with openrouter, Qwen 0.6b and 4B are missing.

Expected result

They should be in the dropdown

Actual result

They are missing

Variations tried (optional)

No response

App Version

v3.30.1

API Provider (optional)

OpenRouter

Model Used (optional)

No response

Roo Code Task Links (optional)

No response

Relevant logs or errors (optional)

[roomote]

I'm exploring a potential fix to add support for the Qwen3 Embedding 0.6B and 4B models for code indexing with OpenRouter. I'll investigate the correct vector sizes and open a PR to share progress.

[dmarkey]

Model Type Models Size Layers Sequence Length Embedding Dimension MRL Support Instruction Aware
Text Embedding Qwen3-Embedding-0.6B 0.6B 28 32K 1024 Yes Yes
Text Embedding Qwen3-Embedding-4B 4B 36 32K 2560 Yes Yes
Text Embedding Qwen3-Embedding-8B 8B 36 32K 4096 Yes Yes

@roomote

[roomote]

Thanks @dmarkey for providing the embedding dimensions! I see you need:

• Qwen3-Embedding-0.6B: 1024 dimensions
• Qwen3-Embedding-4B: 2560 dimensions
• Qwen3-Embedding-8B: 4096 dimensions

I'll explore the codebase to add support for these models in the OpenRouter provider for code indexing.

[roomote]

@dmarkey I've successfully implemented support for the Qwen3 embedding models you requested!

✅ PR Created: #9060
🟢 All CI checks passed

The implementation adds:

• qwen/qwen3-embedding-0.6b with 1024 dimensions
• qwen/qwen3-embedding-4b with 2560 dimensions
• qwen/qwen3-embedding-8b with 4096 dimensions (was already present)

These models are now available in the OpenRouter provider dropdown for code indexing. The PR is ready for review!