qwen
Qwen: Qwen3 Coder 30B A3B Instruct
qwen/qwen3-coder-30b-a3b-instruct
Our take
AI-assisted analysis based on the sources below. Suitability is an editorial assessment; test it on your own workload.
Qwen3-Coder-30B-A3B-Instruct is an instruction-tuned, text-only language model from Qwen aimed at code generation, repository-scale understanding, and agentic tool use. Its Mixture-of-Experts design activates about 3.3B of 30.5B total parameters per forward pass across 128 experts. The vendor reports a 262,144-token native context, extendable to 1M via Yarn, and a purpose-built function-calling format for agentic coding platforms such as Qwen Code and CLINE. It runs only in non-thinking mode and emits no reasoning blocks. The model card lists an Apache-2.0 license and a June 2025 knowledge cutoff.[1][2]
Strengths
- Sparse MoE design: the card describes 30.5B total parameters with about 3.3B activated per pass, targeting efficiency alongside capacity.[1][2]
- Long context for code: vendor-reported 262,144-token native window, extendable to 1M tokens with Yarn, aimed at repository-scale inputs.[1][2]
- Tool use built in: function calling with a dedicated format, plus tools, tool_choice, and structured_outputs parameters in the OpenRouter listing.[1][2]
Limitations
- Non-thinking mode only: the model does not generate reasoning blocks, and OpenRouter lists its reasoning capability as null.[1][2]
- Text in, text out only; the sources describe no image or other modality support.[1][2]
- Evidence gaps: this pack contains only the OpenRouter catalog and Qwen's own model card, so performance claims are vendor-reported; independent benchmarks, hardware requirements, and language coverage are not detailed and warrant validation on real workloads.[1][2]
Where to use it
- Worth evaluating for agentic coding assistants and tool-calling workflows, where its function-call format and tool parameters may fit.[1][2]
- A candidate for repository-scale code comprehension or refactoring that benefits from very long context; confirm quality on your own codebase first.[1][2]
- Plausible for self-hosted or local deployment via transformers, Ollama, LM Studio, llama.cpp, MLX-LM, or KTransformers; the card requires a recent transformers version and suggests reducing context length if memory is constrained.[2]
Sources
- [1] OpenRouter catalog · Qwen: Qwen3 Coder 30B A3B InstructSources checked: Nov 13, 2026
- [2] Model documentation · Qwen: Qwen3 Coder 30B A3B InstructSources checked: Nov 13, 2026
About this model
Description below is the original text supplied by OpenRouter, shown in the source language.
Qwen3-Coder-30B-A3B-Instruct is a 30.5B parameter Mixture-of-Experts (MoE) model with 128 experts (8 active per forward pass), designed for advanced code generation, repository-scale understanding, and agentic tool use. Built on the...
Specifications
- Model ID
- qwen/qwen3-coder-30b-a3b-instruct
- Creator
- qwen
- Added to OpenRouter
- 31.07.2025
- Context length
- 262,144
- Modality
- text->text
- Input modalities
- text
- Output modalities
- text
- Token pricing
- USD per 1M tokens: in $0.150 / out $0.600
- Knowledge cutoff
- 30.06.2025
- Expiration date
- —
Tasks and capabilities
A grade appears only where the source supplies a matching measurement for that task, the model is active and the cohort holds at least 10 comparable models. The capability column lists facts the source declares — supported formats, reasoning and tool parameters — which are not measured quality.
| Task | Grade | Source-declared capability |
|---|---|---|
| General quality | No measurement | — |
| Coding | No measurement | — |
| Agents | No measurement | parameters: structured_outputs, tool_choice, tools |
| Writing | No measurement | output: text |
| Reasoning | No measurement | — |
| Document analysis | No measurement | — |
| Vision | No measurement | — |
| Image generation | No measurement | — |
| Audio | No measurement | — |
| Video | No measurement | — |
| Embeddings | No measurement | — |
Evaluations
| Source | Benchmark | Value | Grade | Data retrieved |
|---|---|---|---|---|
| Design Arena | models · datavizwin rate 54.7% · rank 107 | 1,085Elo | CC — below the B band | 13.11.2026 |
| Design Arena | models · uicomponentwin rate 54.1% · rank 115 | 1,037Elo | CC — below the B band | 13.11.2026 |
| Design Arena | models · websitewin rate 57.1% · rank 117 | 1,080Elo | CC — below the B band | 13.11.2026 |
How grades work
Each letter grade compares a model only against other models scored by the same source and metric in the current snapshot. S means the top 10%, A the next 20%, B the next 30%, C the rest. Models without a published measurement are shown as not rated — a supported format alone never implies measured quality.
- S — S — top 10% of the cohort
- A — A — next 20% of the cohort
- B — B — next 30% of the cohort
- C — C — below the B band
Providers
Optional live provider statistics from OpenRouter. Missing values are unknown, not zero.
| Provider | Context length | Price in /1M | Price out /1M | Latency p50 (30m, s) | Throughput p50 (30m, tok/s) | Uptime (30m) |
|---|---|---|---|---|---|---|
| SiliconFlowfp8 | 262,144 | $0.070 | $0.280 | — | — | 100% |
| Amazon Bedrock | 262,144 | $0.150 | $0.600 | — | — | 100% |
Provider data retrieved: 13.11.2026
Related articles
No published articles identify this model yet.