Qwen3 14B
Alibaba Qwen3 14B is a dense causal language model with thinking and non-thinking modes, 32K native context window, YaRN scaling, and tool use support.
Model Overview
Capabilities, design details, and architectural traits
Qwen3 14B - dense causal language model
Qwen3 14B is a dense causal language model in the Qwen3 series. Its documented identity centers on switching between a thinking mode and a non-thinking mode, with support for instruction-following, tool use, creative writing, and multilingual tasks.
Documented characteristics
- Thinking / non-thinking modes for different task styles.
- Tool use support.
- Instruction-following behavior.
- 32K native context, with extension to 131K tokens using YaRN-based scaling.
- Support for 100+ languages and dialects.
- Designed for math, programming, and logical inference in thinking mode.
- Designed for general-purpose conversation in non-thinking mode.
The model is positioned as a dense 14B-class Qwen3 model rather than a mixture-of-experts variant.
Benchmark Performance
Independent evaluations · Artificial Analysis
Accuracy & Capability Details
Compare Models Side-by-Side
Evaluate specifications, pricing, and independent benchmark indices
| Model Details | |||
|---|---|---|---|
| General Info | |||
| Provider | Alibaba | Anthropic | Anthropic |
| Release Date | April 28, 2025 | September 22, 2026 | September 28, 2026 |
| Knowledge Cutoff | Mar 2025 | - | Jun 2026 |
| Context & Limits | |||
| Context Window | 33K | 1M Best Context Window | 1M Best Context Window |
| Pricing (per 1M tokens) | |||
| Input Pricing | $0.12 Best Input Pricing | $4 | $2 |
| Output Pricing | $0.24 Best Output Pricing | $20 | $10 |
| Modalities | |||
| Inputs | text | textimagefile | textimagefile |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 6.7 | 57.6 Best Intelligence Index | 56.0 |
| Coding Index | - | - | - |
| Agentic Index | - | - | - |
Humanity's Last Exam
Extremely difficult logical reasoning and knowledge.
Long Context Reasoning
Logical reasoning over long context windows.
SciCode Benchmark
Scientific coding and mathematical modeling.
Independent evaluation data provided by Artificial Analysis. To view the latest benchmarks and full details, visit their official site.
Explore more from Alibaba
Other models by Alibaba
Qwen3.8 Max
Qwen3.8 2.4T A95B
Qwen3.8-Flash-Next
Qwen3.8 27B
Top AI Models
Leading alternatives by intelligence score
