# Qwen — full timeline

69 tracked events from 2024-08-09 to 2026-08-22.

## 2026

- **2026-08-22** `—` sig 3 — [Artificial Analysis evaluates Qwen3.8 27B (medium)](https://changelogs.info/qwen/artificial-analysis-evaluates-qwen3-8-27b-medium)
- **2026-08-22** `—` sig 3 — [Artificial Analysis evaluates Qwen3.8 27B (low)](https://changelogs.info/qwen/artificial-analysis-evaluates-qwen3-8-27b-low)
- **2026-08-21** `—` sig 2 — [OpenRouter catalog: new listing `qwen/qwen3.8-2.4t-a95b`, 1M context](https://changelogs.info/qwen/openrouter-catalog-new-listing-qwen-qwen3-8-2-4t-a95b-1m-context)
- **2026-08-21** `—` sig 2 — [OpenRouter catalog: new listing `qwen/qwen3.8-27b`, 1M context](https://changelogs.info/qwen/openrouter-catalog-new-listing-qwen-qwen3-8-27b-1m-context)
- **2026-08-13** `—` sig 1 — [Qwen3.8-27B FP8 quantized model published](https://changelogs.info/qwen/qwen3-8-27b-fp8-quantized-model-published)
- **2026-08-09** `—` sig 1 — [changelogs.info now tracks Qwen](https://changelogs.info/qwen/coverage)
- **2026-08-08** `—` sig 1 — [Qwen3.8-2.4T-A95B MoE model published](https://changelogs.info/qwen/qwen3-8-2-4t-a95b-moe-model-published)
- **2026-08-08** `—` sig 1 — [Qwen3.8-2.4T-A95B FP8 quantized weights published](https://changelogs.info/qwen/qwen3-8-2-4t-a95b-fp8-quantized-weights-published)
- **2026-08-05** `—` sig 1 — [Qwen publishes Qwen3.8-27B on Hugging Face](https://changelogs.info/qwen/qwen-publishes-qwen3-8-27b-on-hugging-face)
- **2026-06-26** `—` sig 1 — [Qwen publishes Qwen3-ForcedAligner-0.6B on Hugging Face](https://changelogs.info/qwen/qwen-publishes-qwen3-forcedaligner-0-6b-on-hugging-face)
- **2026-06-26** `—` sig 1 — [Qwen publishes Qwen3-ASR-0.6B on Hugging Face](https://changelogs.info/qwen/qwen-publishes-qwen3-asr-0-6b-on-hugging-face)
- **2026-06-26** `—` sig 1 — [Qwen publishes Qwen3-ASR-1.7B on Hugging Face](https://changelogs.info/qwen/qwen-publishes-qwen3-asr-1-7b-on-hugging-face)
- **2026-06-22** `—` sig 1 — [Qwen publishes AgentWorld-35B-A3B on Hugging Face](https://changelogs.info/qwen/qwen-publishes-agentworld-35b-a3b-on-hugging-face)
- **2026-05-21** `—` sig 1 — [Qwen publishes Qwen-Image-Bench evaluation model](https://changelogs.info/qwen/qwen-publishes-qwen-image-bench-evaluation-model)
- **2026-04-27** `—` sig 1 — [Qwen publishes SAE residual stream weights for Qwen3.5-35B-A3B-Base (W128K, L0 100)](https://changelogs.info/qwen/qwen-publishes-sae-residual-stream-weights-for-qwen3-5-35b-a3b-base-w128k-l0-100)
- **2026-04-27** `—` sig 1 — [Qwen publishes SAE residual stream weights for Qwen3.5-35B-A3B-Base (W32K, L0 50)](https://changelogs.info/qwen/qwen-publishes-sae-residual-stream-weights-for-qwen3-5-35b-a3b-base-w32k-l0-50)
- **2026-04-27** `—` sig 1 — [Qwen publishes SAE residual stream weights for Qwen3-30B-A3B-Base (W128K, L0 100)](https://changelogs.info/qwen/qwen-publishes-sae-residual-stream-weights-for-qwen3-30b-a3b-base-w128k-l0-100)
- **2026-04-27** `—` sig 1 — [Qwen publishes SAE residual stream weights for Qwen3-30B-A3B-Base (W32K, L0 50)](https://changelogs.info/qwen/qwen-publishes-sae-residual-stream-weights-for-qwen3-30b-a3b-base-w32k-l0-50)
- **2026-04-27** `—` sig 1 — [Qwen publishes SAE residual stream weights for Qwen3.5-27B (W80K, L0 100)](https://changelogs.info/qwen/qwen-publishes-sae-residual-stream-weights-for-qwen3-5-27b-w80k-l0-100)
- **2026-04-27** `—` sig 1 — [Qwen publishes SAE residual stream weights for Qwen3.5-27B (W80K, L0 50)](https://changelogs.info/qwen/qwen-publishes-sae-residual-stream-weights-for-qwen3-5-27b-w80k-l0-50)
- **2026-04-27** `—` sig 1 — [Qwen publishes residual-stream SAE for Qwen3.5-9B-Base (64K width, L0=100)](https://changelogs.info/qwen/qwen-publishes-residual-stream-sae-for-qwen3-5-9b-base-64k-width-l0-100)
- **2026-04-27** `—` sig 1 — [Qwen publishes SAE residual stream weights for Qwen3.5-2B-Base (W32K, L0 100)](https://changelogs.info/qwen/qwen-publishes-sae-residual-stream-weights-for-qwen3-5-2b-base-w32k-l0-100)
- **2026-04-27** `—` sig 1 — [Qwen publishes residual-stream SAE for Qwen3.5-9B-Base (64K width, L0=50)](https://changelogs.info/qwen/qwen-publishes-residual-stream-sae-for-qwen3-5-9b-base-64k-width-l0-50)
- **2026-04-27** `—` sig 1 — [Qwen publishes SAE residual stream weights for Qwen3.5-2B-Base (W32K, L0 50)](https://changelogs.info/qwen/qwen-publishes-sae-residual-stream-weights-for-qwen3-5-2b-base-w32k-l0-50)
- **2026-04-27** `—` sig 1 — [Qwen publishes residual-stream SAE for Qwen3-8B-Base (64K width, L0=100)](https://changelogs.info/qwen/qwen-publishes-residual-stream-sae-for-qwen3-8b-base-64k-width-l0-100)
- **2026-04-27** `—` sig 1 — [Qwen publishes residual-stream SAE for Qwen3-8B-Base (64K width, L0=50)](https://changelogs.info/qwen/qwen-publishes-residual-stream-sae-for-qwen3-8b-base-64k-width-l0-50)
- **2026-04-27** `—` sig 1 — [Qwen: sparse autoencoder research model repository published](https://changelogs.info/qwen/qwen-sparse-autoencoder-research-model-repository-published)
- **2026-04-27** `—` sig 1 — [Qwen: sparse autoencoder research model repository published](https://changelogs.info/qwen/qwen-sparse-autoencoder-research-model-repository-published)
- **2026-04-15** `—` sig 1 — [Qwen3.6-35B-A3B MoE multimodal model published](https://changelogs.info/qwen/qwen3-6-35b-a3b-moe-multimodal-model-published)
- **2026-03-03** `—` sig 1 — [Qwen3.5-397B-A17B MoE GPTQ Int4 quantized variant published on Hugging Face](https://changelogs.info/qwen/qwen3-5-397b-a17b-moe-gptq-int4-quantized-variant-published-on-hugging-face)
- **2026-03-03** `—` sig 1 — [Qwen3.5-122B-A10B MoE GPTQ Int4 quantized variant published on Hugging Face](https://changelogs.info/qwen/qwen3-5-122b-a10b-moe-gptq-int4-quantized-variant-published-on-hugging-face)
- **2026-02-25** `—` sig 1 — [Qwen3.5-122B-A10B MoE FP8 quantized variant published on Hugging Face](https://changelogs.info/qwen/qwen3-5-122b-a10b-moe-fp8-quantized-variant-published-on-hugging-face)
- **2026-02-24** `—` sig 1 — [Qwen publishes Qwen3.5-122B-A10B on Hugging Face](https://changelogs.info/qwen/qwen-publishes-qwen3-5-122b-a10b-on-hugging-face)
- **2026-02-24** `—` sig 1 — [Qwen publishes Qwen3.5-27B on Hugging Face](https://changelogs.info/qwen/qwen-publishes-qwen3-5-27b-on-hugging-face)
- **2026-02-18** `—` sig 1 — [Qwen publishes FP8 quantized Qwen3.5-397B-A17B](https://changelogs.info/qwen/qwen-publishes-fp8-quantized-qwen3-5-397b-a17b)
- **2026-02-16** `—` sig 1 — [Qwen publishes Qwen3.5-397B-A17B on Hugging Face](https://changelogs.info/qwen/qwen-publishes-qwen3-5-397b-a17b-on-hugging-face)
- **2026-02-13** `—` sig 1 — [Qwen publishes WebWorld-32B web agent world model](https://changelogs.info/qwen/qwen-publishes-webworld-32b-web-agent-world-model)
- **2026-02-13** `—` sig 1 — [Qwen publishes WebWorld-14B web agent world model](https://changelogs.info/qwen/qwen-publishes-webworld-14b-web-agent-world-model)
- **2026-02-13** `—` sig 1 — [Qwen publishes WebWorld-8B web-agent simulator model](https://changelogs.info/qwen/qwen-publishes-webworld-8b-web-agent-simulator-model)

## 2025

- **2025-09-22** `—` sig 3 — [Qwen3Guard: first safety guardrail model in Qwen family](https://changelogs.info/qwen/qwen3guard-first-safety-guardrail-model-in-qwen-family)
- **2025-08-18** `—` sig 3 — [Qwen-Image-Edit: image editing model with text rendering](https://changelogs.info/qwen/qwen-image-edit-image-editing-model-with-text-rendering)
- **2025-08-04** `—` sig 3 — [Qwen-Image: 20B image foundation model with native text rendering](https://changelogs.info/qwen/qwen-image-20b-image-foundation-model-with-native-text-rendering)
- **2025-07-27** `—` sig 3 — [GSPO: Group Sequence Policy Optimization for stable RL scaling](https://changelogs.info/qwen/gspo-group-sequence-policy-optimization-for-stable-rl-scaling)
- **2025-07-24** `—` sig 3 — [Qwen-MT: updated translation API supports 92 languages](https://changelogs.info/qwen/qwen-mt-updated-translation-api-supports-92-languages)
- **2025-07-22** `—` sig 4 — [Qwen3-Coder-480B-A35B: open MoE model for agentic coding](https://changelogs.info/qwen/qwen3-coder-480b-a35b-open-moe-model-for-agentic-coding)
- **2025-06-27** `—` sig 3 — [Qwen-TTS adds three Chinese dialects](https://changelogs.info/qwen/qwen-tts-adds-three-chinese-dialects)
- **2025-06-26** `—` sig 3 — [Qwen VLo: unified multimodal understanding and generation model](https://changelogs.info/qwen/qwen-vlo-unified-multimodal-understanding-and-generation-model)
- **2025-06-05** `—` sig 3 — [Qwen3 Embedding series: text embedding and reranking models](https://changelogs.info/qwen/qwen3-embedding-series-text-embedding-and-reranking-models)
- **2025-04-28** `—` sig 5 — [Qwen3: new frontier model family released](https://changelogs.info/qwen/qwen3-new-frontier-model-family-released)
- **2025-03-27** `—` sig 3 — [Qwen releases QVQ-Max visual reasoning model](https://changelogs.info/qwen/qwen-releases-qvq-max-visual-reasoning-model)
- **2025-03-26** `—` sig 4 — [Qwen releases Qwen2.5-Omni end-to-end multimodal model](https://changelogs.info/qwen/qwen-releases-qwen2-5-omni-end-to-end-multimodal-model)
- **2025-03-23** `—` sig 3 — [Qwen open-sources Qwen2.5-VL-32B-Instruct under Apache 2.0](https://changelogs.info/qwen/qwen-open-sources-qwen2-5-vl-32b-instruct-under-apache-2-0)
- **2025-03-05** `—` sig 3 — [Qwen publishes QwQ-32B with RL scaling research](https://changelogs.info/qwen/qwen-publishes-qwq-32b-with-rl-scaling-research)
- **2025-02-24** `—` sig 1 — [Qwen publishes blog for QwQ-Max-Preview with no substantive content](https://changelogs.info/qwen/qwen-publishes-blog-for-qwq-max-preview-with-no-substantive-content)
- **2025-01-28** `—` sig 3 — [Qwen publishes Qwen2.5-Max MoE model research blog](https://changelogs.info/qwen/qwen-publishes-qwen2-5-max-moe-model-research-blog)
- **2025-01-26** `—` sig 4 — [Qwen open-sources Qwen2.5-1M models with 1M-token context windows](https://changelogs.info/qwen/qwen-open-sources-qwen2-5-1m-models-with-1m-token-context-windows)
- **2025-01-26** `—` sig 4 — [Qwen releases Qwen2.5-VL vision-language model family](https://changelogs.info/qwen/qwen-releases-qwen2-5-vl-vision-language-model-family)
- **2025-01-20** `—` sig 1 — [Qwen proposes global-batch load balancing for MoE training](https://changelogs.info/qwen/qwen-proposes-global-batch-load-balancing-for-moe-training)
- **2025-01-13** `—` sig 1 — [Qwen research on process supervision for mathematical reasoning](https://changelogs.info/qwen/qwen-research-on-process-supervision-for-mathematical-reasoning)

## 2024

- **2024-12-24** `—` sig 3 — [Qwen announces QVQ visual reasoning model](https://changelogs.info/qwen/qwen-announces-qvq-visual-reasoning-model)
- **2024-11-27** `—` sig 3 — [Qwen announces QwQ reasoning model](https://changelogs.info/qwen/qwen-announces-qwq-reasoning-model)
- **2024-11-14** `—` sig 4 — [Qwen2.5-Turbo extends context window to 1M tokens](https://changelogs.info/qwen/qwen2-5-turbo-extends-context-window-to-1m-tokens)
- **2024-11-11** `—` sig 5 — [Qwen2.5-Coder series open-sourced, 32B Instruct matches GPT-4o on coding](https://changelogs.info/qwen/qwen2-5-coder-series-open-sourced-32b-instruct-matches-gpt-4o-on-coding)
- **2024-09-18** `—` sig 5 — [Qwen2.5 foundation model family announced](https://changelogs.info/qwen/qwen2-5-foundation-model-family-announced)
- **2024-09-18** `—` sig 5 — [Qwen2.5 LLM series open-sourced, 0.5B to 72B parameters](https://changelogs.info/qwen/qwen2-5-llm-series-open-sourced-0-5b-to-72b-parameters)
- **2024-09-18** `—` sig 4 — [Qwen2.5-Coder released, CodeQwen rebranded to Qwen-Coder](https://changelogs.info/qwen/qwen2-5-coder-released-codeqwen-rebranded-to-qwen-coder)
- **2024-09-18** `—` sig 4 — [Qwen2.5-Math series open-sourced with reward model](https://changelogs.info/qwen/qwen2-5-math-series-open-sourced-with-reward-model)
- **2024-08-28** `—` sig 4 — [Qwen2-VL: vision-language model with 20+ minute video understanding](https://changelogs.info/qwen/qwen2-vl-vision-language-model-with-20-minute-video-understanding)
- **2024-08-09** `—` sig 3 — [Qwen2-Audio: multi-modal model accepting audio and text inputs](https://changelogs.info/qwen/qwen2-audio-multi-modal-model-accepting-audio-and-text-inputs)

---
Canonical: https://changelogs.info/qwen/timeline