ai/qwen3.5 repository overview

⁠To run Qwen3.5 locally - Read our Guide!⁠

Unsloth Dynamic 2.0⁠ achieves superior accuracy & outperforms other leading quants.

⁠ ⁠ ⁠

Disable thinking via --chat-template-kwargs '{"enable_thinking":false}'. Read our guide⁠.

⁠Mar 5 'Final' Update:

• All GGUFs now use our new imatrix data. See some improvements in chat, coding, long context, and tool-calling use-cases.
• GGUFs now updated with an improved quantization algorithm.
• Rest of variants like Q8_0, Q4_K_M, BF16 are now uploaded.
• See our new benchmarks for 122B-A10B here⁠.

For Qwen3.5-35B-A3B, we primarily reduced the maximum KLD:

⁠Feb 27 Update: GGUFs Refreshed + Tool-calling fixes + Benchmarks

• Qwen3.5 is now updated with improved tool-calling & coding performance! See improvements via Claude Code, Codex.
• We also benchmarked GGUFs & removed MXFP4 layers from 3 quants. Read analysis here.⁠

Please follow the correct instructions / settings in our guide here⁠.

⁠Fine-tuning and RL Qwen3.5

• You can also fine-tune and perform reinforcement learning (RL)⁠ on all Qwen3.5 models with Unsloth via our free Colab notebooks⁠.
• Read our Qwen3.5 fine-tuning guide⁠ for tips, VRAM requirements, code and more here⁠.

⁠Qwen3.5-35B-A3B

Over recent months, we have intensified our focus on developing foundation models that deliver exceptional utility and performance. Qwen3.5 represents a significant leap forward, integrating breakthroughs in multimodal learning, architectural efficiency, reinforcement learning scale, and global accessibility to empower developers and enterprises with unprecedented capability and efficiency.

⁠Qwen3.5 Highlights

Qwen3.5 features the following enhancement:

• Unified Vision-Language Foundation: Early fusion training on multimodal tokens achieves cross-generational parity with Qwen3 and outperforms Qwen3-VL models across reasoning, coding, agents, and visual understanding benchmarks.

• Efficient Hybrid Architecture: Gated Delta Networks combined with sparse Mixture-of-Experts deliver high-throughput inference with minimal latency and cost overhead.

• Scalable RL Generalization: Reinforcement learning scaled across million-agent environments with progressively complex task distributions for robust real-world adaptability.

• Global Linguistic Coverage: Expanded support to 201 languages and dialects, enabling inclusive, worldwide deployment with nuanced cultural and regional understanding.

• Next-Generation Training Infrastructure: Near-100% multimodal training efficiency compared to text-only training and asynchronous RL frameworks supporting massive-scale agent scaffolds and environment orchestration.

For more details, please refer to our blog post Qwen3.5⁠.

⁠Model Overview

• Type: Causal Language Model with Vision Encoder
• Training Stage: Pre-training & Post-training
• Language Model
  • Number of Parameters: 35B in total and 3B activated
  • Hidden Dimension: 2048
  • Token Embedding: 248320 (Padded)
  • Number of Layers: 40
  • Hidden Layout: 10 × (3 × (Gated DeltaNet → MoE) → 1 × (Gated Attention → MoE))
  • Gated DeltaNet:
    • Number of Linear Attention Heads: 32 for V and 16 for QK
    • Head Dimension: 128
  • Gated Attention:
    • Number of Attention Heads: 16 for Q and 2 for KV
    • Head Dimension: 256
    • Rotary Position Embedding Dimension: 64
  • Mixture Of Experts
    • Number of Experts: 256
    • Number of Activated Experts: 8 Routed + 1 Shared
    • Expert Intermediate Dimension: 512
  • LM Output: 248320 (Padded)
  • MTP: trained with multi-steps
• Context Length: 262,144 natively and extensible up to 1,010,000 tokens.

⁠Benchmark Results

⁠Language