Qwen 3.5 122B-A10B
AlibabaLLMs
87.5
Performance
★ 4.4
Rating
98
Reviews
Text GenerationLargeMultimodalReasoningOpen Weight
About
Medium-sized Qwen 3.5 MoE model balancing strong capability with efficient inference for production deployments.
Strengths
Strong tool use and function-calling (BFCL-V4). Good balance of capability and efficiency with 10B active parameters. Open-weight Apache 2.0. Excellent for multilingual and code workflows.
Specifications
- Context window
- 256,000
- Parameters
- 122B
Pricing
- Input cost
- Free
- Output cost
- Free
Open-weight (Apache 2.0). Free to download.
Speed & Latency
- 70
- tokens/sec
- 350ms
- time to first token
Available On
HuggingFaceTogether AIAlibaba Cloud
Features
function callingstreamingjson modesystem messages
Performance Trend
Benchmark score trends over time for the top 5 benchmarks.
Loading history...
Benchmarks
Scores from various benchmark tests; higher is better.
| Test | Score | Percentile | Source |
|---|---|---|---|
| BigBench Hard | 81.5 | p92 | seed |
| DROP | 82.5 | p91 | seed |
| GSM8K | 90.0 | p95 | seed |
| HumanEval | 85.5 | p95 |
