Qwen3.8 27B by Alibaba: 27.78B dense multimodal model, Apache 2.0, 262K context, text-image-video input, strong agentic coding and computer-use.
Capabilities, design details, and architectural traits
Qwen3.8 27B is a dense, locally deployable multimodal model from Alibaba's Qwen team, positioned as the strongest candidate around 30 billion parameters for local coding assistants, bounded agents, and private multimodal workflows. Unlike the 2.4-trillion-parameter Qwen3.8-Max, this checkpoint ships as a single 27.78B dense model with open weights under Apache 2.0.
| Trait | Detail |
|---|---|
| Dense architecture at 27.78B | Not a mixture-of-experts; the full 27.78 billion parameters are active, making it suitable for quantization onto high-end workstations |
| Multimodal input | Accepts text, images, and video in a single checkpoint |
| Apache 2.0 license | Open weights with permissive licensing for local and commercial use |
| 262K native context | 262,144-token native context window; one million tokens requires YaRN scaling, which Qwen warns can hurt shorter prompts |
| Single post-trained checkpoint | No separate Base and Instruct variants published at launch |
The training corpus, token count, knowledge cutoff, and post-training recipe are undisclosed.
Independent evaluations · Artificial Analysis
Evaluate specifications, pricing, and independent benchmark indices
| Model Details | |||
|---|---|---|---|
| General Info | |||
| Provider | Alibaba | Anthropic | Anthropic |
| Release Date | August 14, 2026 | July 24, 2026 | June 9, 2026 |
| Knowledge Cutoff | - | May 2026 | - |
| Context & Limits | |||
| Context Window | 262K | 1M Best Context Window | 1M Best Context Window |
| Pricing (per 1M tokens) | |||
| Input Pricing | $0.50 Best Input Pricing | $5 | $10 |
| Output Pricing | $3 Best Output Pricing | $25 | $50 |
| Modalities | |||
| Inputs | textimagefilevideo | textimage | textimagefile |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 52.0 | 63.1 Best Intelligence Index | 62.1 |
| Coding Index | 68.1 | 78.0 Best Coding Index | 76.5 |
| Agentic Index | 50.9 | 59.2 Best Agentic Index | 56.6 |
Graduate-level reasoning and expert Q&A evaluation.
Extremely difficult logical reasoning and knowledge.
Logical reasoning over long context windows.
Independent evaluation data provided by Artificial Analysis. To view the latest benchmarks and full details, visit their official site.
Prices in USD.
