Qwen3 30B A3B Instruct 2507
Alibaba · Released Jul 2025
A 30.5-billion-parameter mixture-of-experts model from Alibaba that activates 3.3 billion parameters per token with a 128k-token context window.
- Strengths
- Balances efficiency and capability through sparse activation, making it suitable for tasks that don't require the full model weight while maintaining reasonable quality.
- Best for
- General-purpose instruction-following where the 3.3B active parameters are sufficient and latency or throughput is a concern relative to larger models.
- Limitations
- The 3.3B active parameter count limits performance on complex reasoning and knowledge-intensive tasks compared to denser models like Qwen3 Next 80B A3B.
Input / 1M
$0.09
Output / 1M
$0.3
Cached input / 1M
—
Context window
262K
Price history
Input (solid)Output (dashed)
Price change
- 30d
- in increased 86.9% out increased 55.4%
- 90d
- in increased 86.9% out increased 55.4%
- 1y
- in increased 86.9% out increased 55.4%
- Since launch
- in increased 86.9% out increased 55.4%
Snapshots
| Effective | Input | Output | Cached in | Note | Source |
|---|---|---|---|---|---|
| 11 Sep 2026 | $0.09 | $0.3 | — | Imported from OpenRouter | openrouter.ai |
| 10 Sep 2026 | $0.0482 | $0.193 | — | Imported from OpenRouter | openrouter.ai |
| 25 Jul 2026 | $0.0482 | $0.193 | — | Imported from OpenRouter | openrouter.ai |
| 15 Jul 2026 | $0.1 | $0.3 | — | Imported from OpenRouter | openrouter.ai |
| 11 Jun 2026 | $0.0482 | $0.193 | — | Imported from OpenRouter | openrouter.ai |