Articles tagged "Qwen"

Qwen-Image-2.1, Alibaba's open model (Sep 20, 2026), unifies 7B gen and editing with transparent PNG, 10-image edits. VRAM unofficial; non-commercial license.

Qwen3.8-Omni-Flash: Alibaba's omni-modal model, 1M-token context, $0.15/million input. Weights closed, no local run; text-only output, unlike its predecessor.

As of Sep 2026, Perplexity's Hybrid Compute feature for the Mac Computer agent starts tasks in the cloud and hands off only the steps involving sensitive data to a local model (PPLX Qwen 3.8 27B) on Apple Silicon. This article explains, based on primary sources, how the local/cloud boundary is decided, the hardware requirements, and how it differs from Apple's Private Cloud Compute and fully local setups.

**[Software Mansion's React Native ExecuTorch](https://docs.swmansion.com/react-native-executorch/) shipped a `useLLM` hook** that gives React Native apps **native on-device LLM integration**. **Supported models**: quantized Qwen (2.5 / 3 / 3.5), Llama 3.2, Hammer 2.1, Phi 4 Mini, SmolLM 2, LFM2.5 (vision), and **[Gemma 4](../columns/gemma-4-technical-report-2026-07) (vision + audio)**. **Two operating modes**: **Functional/Stateless** (developers manage conversation history via `generate()` + `response`; tool calling and chat config don't apply) and **Managed/Stateful** (`sendMessage()` maintains conversation state, parses tool calls, and runs callbacks automatically). **Key features**: token batching (groups tokens before re-render), tool calling (model invokes external functions via tool schemas with automatic parsing and callbacks), vision-language / audio multimodal inputs, generation control (temperature / top-p / repetition penalty / mid-stream interruption), and **JSONSchema- or Zod-backed structured output**. **Use cases**: on-device chatbots without server dependencies, privacy-first conversation UIs, in-app function calling (calendar events, flashlight, etc.), multimodal features (image analysis, audio transcription). **Positioning**: implementation-level evidence that **cutting-edge open-weight LLMs like [Gemma 4's encoder-free 12B](../columns/gemma-4-technical-report-2026-07) and [Qwen 3.6-35B](../columns/qwen36-35b-a3b-uncensored-abliterated-2026-07) natively fit into iOS / Android apps** — a major deliverable from the React Native ecosystem.

A 2026 Q2 performance benchmark of Gemma 4 (E2B / E4B / 26B MoE / 31B Dense) against the major open-weights peers — Llama 4, Qwen 3.5, Mistral, and DeepSeek — across MMLU-Pro, GPQA, HumanEval, MATH-500, and MT-Bench. Adds throughput (tokens / s), memory efficiency (quality per GB VRAM), cost per million tokens, Japanese-language performance, native function calling, and Apache 2.0 / MIT / commercial-use licensing as of May 2026, plus a use-case selection matrix for in-house LLM, edge AI, coding assistants, and RAG.