Supported models

What hf2q converts and serves.

See which model families hf2q converts and serves—and which capabilities each one exposes.

01

Model families

Each family lives in its own module inside hf2q's arch-registry — the single source of truth for tensor catalogs, quality thresholds, and smoke prompts.

FamilyArchitectureServesNotes
Gemma 4Dense + MoEChat completions with tools, streaming, and grammar-constrained samplingLauncher: scripts/serve_gemma4_opencode.sh, default port 8082
Qwen 3.5 / 3.6 / 3.8Dense + MoE + multi-token predictionChat completions with tools and streaming; image inputs with a bound vision projector (serve --mmproj)Launchers: scripts/serve_qwen36_opencode.sh and serve_qwen38_opencode.sh (3.8-27B multimodal), default port 8081. Convert disk floor: 100 GB dense, 150 GB MoE
DeepSeek-V4-Flash-0731Compressed-attention MoEChat completions with tools and streamingLauncher: scripts/serve_deepseek4_opencode.sh, default port 8081. Run one large family at a time. Embeddings and image inputs are refused explicitly
BERT / Nomic-BERTEncoder, embedding-onlyEmbeddings via /v1/embeddings
02

Quantization

hf2q convert accepts two families of --quant values. Anything else fails with a typed, actionable error.

  • Standard llama.cpp ftypes

    All supported families

    f32f16bf16q4_0q4_1q5_0q5_1q8_0q2_kq3_k_sq3_k_mq3_k_lq4_k_sq4_k_mq5_k_sq5_k_mq6_kiq4_nl

    Byte-identical to stock llama-quantize output for the same ftype.

  • APEX algorithmic tiers

    MoE architectures only

    apex-qualityapex-i-qualityapex-balancedapex-i-balancedapex-compactapex-i-compactapex-mini

    I-tier variants require imatrix data via --imatrix or --imatrix-corpus.

    Reserved names (dwq, unqualified apex, tq1_0 / tq2_0) fail with typed, actionable errors.

03

The support boundary

What the supported surface requires, and where it ends.

  • Platform

    macOS on Apple Silicon (M1 or newer). The inference path is Metal-only, and the Rust toolchain is pinned at 1.88.0.

  • Disk floors

    Conversion enforces per-architecture disk floors — 100 GB for Qwen 3.5 dense, 150 GB for Qwen 3.5 MoE. Preflight refuses to start below the floor plus 10 GB.

  • Fail-closed support

    The published support contract is limited to families on this page, and reserved quant names fail with typed errors.

  • Serving surface

    One OpenAI-compatible endpoint per launched family: /v1/chat/completions, /v1/embeddings, and /v1/models, with streaming SSE and a persistent block-prefix KV cache.