Qwen3 30B A3B Thinking 2507
Alibaba · Released Aug 2025
A 30.5-billion-parameter mixture-of-experts model from Alibaba that activates 3.3 billion parameters per token and uses extended reasoning capabilities.
- Strengths
- Delivers reasoning-focused task performance through chain-of-thought processing while maintaining efficient token usage through sparse activation.
- Best for
- Complex problem-solving, multi-step reasoning, and decision-making tasks where model transparency or intermediate reasoning steps are valuable.
- Limitations
- Smaller than the 235B and 480B thinking models in the Qwen3 lineup, with an 81k-token context window that is narrower than many contemporary models.
Input / 1M
$0.2
Output / 1M
$2.40
Cached input / 1M
—
Context window
81K
Price history
Input (solid)Output (dashed)
Price change
- 30d
- in — out —
- 90d
- in increased 150.0% out increased 500.0%
- 1y
- in increased 150.0% out increased 500.0%
- Since launch
- in increased 150.0% out increased 500.0%
Snapshots
| Effective | Input | Output | Cached in | Note | Source |
|---|---|---|---|---|---|
| 29 Jul 2026 | $0.2 | $2.40 | — | Imported from OpenRouter | openrouter.ai |
| 2 Jul 2026 | $0.13 | $1.56 | — | Imported from OpenRouter | openrouter.ai |
| 11 Jun 2026 | $0.08 | $0.4 | $0.08 | Imported from OpenRouter | openrouter.ai |