Qwen3 8B
Alibaba · Released Apr 2025
An 8-billion-parameter dense language model from Alibaba with a 131k-token context window.
- Strengths
- Offers a smaller parameter footprint than the 14B and 32B dense models in Alibaba's lineup, making it faster and cheaper to run while maintaining reasonable quality across general tasks.
- Best for
- Applications prioritizing inference speed and resource efficiency, such as real-time chat, embedded systems, or high-volume batch processing where full context capacity matters more than scale.
- Limitations
- Will underperform on complex reasoning, code generation, and specialized tasks compared to Alibaba's larger models like the Qwen3 32B or mixture-of-experts variants.
Input / 1M
$0.117
Output / 1M
$0.455
Cached input / 1M
—
Context window
131K
Price history
Input (solid)Output (dashed)
Price change
- 30d
- in — out —
- 90d
- in — out —
- 1y
- in increased 134.0% out increased 13.8%
- Since launch
- in increased 134.0% out increased 13.8%
Snapshots
| Effective | Input | Output | Cached in | Note | Source |
|---|---|---|---|---|---|
| 2 Jul 2026 | $0.117 | $0.455 | — | Imported from OpenRouter | openrouter.ai |
| 11 Jun 2026 | $0.05 | $0.4 | $0.05 | Imported from OpenRouter | openrouter.ai |