Qwen3 Coder 30B A3B Instruct

Tool callingStructured output

Hugging Face Launch post Website Docs

Smaller Qwen coder for efficient local agents and repo-level fixes

Context: 262K
Released: 1y ago
Parameters: 30B
Input / 1M tokens: $0.07
Output / 1M tokens: $0.28

Sources:models.devOpenRouter

Benchmarks

DeepSWE is the primary coding benchmark where an exact model evaluation exists. Other sourced evaluations appear below; missing scores are not zero.

Not evaluated in DeepSWE 1.1 as of 2026-09-22. Results are only shown for exact model versions in the official dataset. Official leaderboard.

No benchmark results have been matched to this model yet.

Specifications

Developer: Qwen
Model family: Qwen
Type: Language model
Predecessor: —
Successor: —
Released: April 1, 2025
Model ID: qwen/qwen3-coder-30b-a3b-instruct
Openness: Open weights
License: —
Parameters: 30B
Input modalities: text
Output modalities: text
Modality: text->text
Native context: 256K
Tokenizer: Qwen3
Max output tokens: 236K
Knowledge cutoff: 2025-04
Hugging Face: Qwen/Qwen3-Coder-30B-A3B-Instruct
Source freshness: Metadata synced Oct 8, 2026, 10:17 PM UTC

Open-weight builds

Ollama variants

Every published Ollama tag matched to this model, including parameter editions, instruction/base variants, and quantized builds.

6 builds

Variant	Parameters	Type	Quantization	Size	Context	Input
qwen3-coder:30b-a3b-fp16	30B	—	FP16	61 GB	256K	text
qwen3-coder:latest	30B	—	Q4_K_M	19 GB	256K	text
qwen3-coder:30b-a3b-q8_0	30B	—	Q8_0	32 GB	256K	text
qwen3-coder:480b-a35b-fp16	480B	—	FP16	960 GB	256K	text
qwen3-coder:480b	480B	—	Q4_K_M	290 GB	256K	text
qwen3-coder:480b-a35b-q8_0	480B	—	Q8_0	510 GB	256K	text

Additional source evidence

Source model: alibaba/qwen3-coder-30b-a3b-instruct
Source family: qwen
Last updated: 2025-04
Knowledge cutoff: 2025-04
Model context: 262K
Model output limit: 66K
Open weights: Yes
File attachments: No
Reasoning: No
Tool calling: Yes
Temperature control: Yes

Downloadable weights

Hugging Face

Published benchmarks

Benchmark	Score	Evaluation details	Evidence
Artificial Analysis Coding Index	19.4 · index	2026-06-02	Source
SciCode	27.8 · percent correct	2026-06-02	Source
Terminal-Bench Hard	15.2 · success rate	2026-06-02	Source

Complete source record (JSON)

{
  "id": "alibaba/qwen3-coder-30b-a3b-instruct",
  "name": "Qwen3-Coder 30B-A3B Instruct",
  "description": "Smaller Qwen coder for efficient local agents and repo-level fixes",
  "family": "qwen",
  "last_updated": "2025-04",
  "knowledge": "2025-04",
  "attachment": false,
  "reasoning": false,
  "tool_call": true,
  "temperature": true,
  "modalities": {
    "input": [
      "text"
    ],
    "output": [
      "text"
    ]
  },
  "limit": {
    "context": 262144,
    "output": 65536
  },
  "benchmarks": [
    {
      "name": "Artificial Analysis Coding Index",
      "score": 19.4,
      "source": "https://openrouter.ai/qwen/qwen3-coder-30b-a3b-instruct/benchmarks",
      "metric": "index",
      "date": "2026-06-02"
    },
    {
      "name": "SciCode",
      "score": 27.8,
      "source": "https://openrouter.ai/qwen/qwen3-coder-30b-a3b-instruct/benchmarks",
      "metric": "percent correct",
      "date": "2026-06-02"
    },
    {
      "name": "Terminal-Bench Hard",
      "score": 15.2,
      "source": "https://openrouter.ai/qwen/qwen3-coder-30b-a3b-instruct/benchmarks",
      "metric": "success rate",
      "date": "2026-06-02"
    }
  ],
  "release_date": "2025-04",
  "open_weights": true,
  "weights": [
    {
      "label": "Hugging Face",
      "url": "https://huggingface.co/Qwen/Qwen3-Coder-30B-A3B-Instruct"
    }
  ]
}

Infrastructure

Provider availability

All providers

Provider	Endpoint	Context	Input /M tokens	Output /M tokens	Cached input /M tokens	Cache write /M tokens	Uptime 1d	Data handling
SiliconFlow	siliconflow/fp8	262K	$0.07	$0.28	—	—	100.00%	ZDR: Yes Training: Unknown Prompt logging: Unknown Retention: Unknown ZDR source Privacy policy Checked 2026-10-15

Provider data:OpenRouter

Deployment graph

Can I run this?

Vertical Tabs only shows compatibility when it has structured evidence tying this exact model to hardware, runtime, quantization, memory use, and observed performance.

Browse hardware