Reference · rapid-mlx 0.12.18 · ← Back to README

Model aliases

Rapid-MLX ships 175 text aliases plus a separate audio registry (kokoro / whisper / parakeet / vibevoice / voxcpm / chatterbox / dia). An alias is a short, stable name that resolves to a canonical HuggingFace MLX repo and auto-applies the right tool-call parser, reasoning parser, hybrid-cache flag, MoE flag, spec-decode eligibility, and PFlash tier.

Source of truth. The text-alias registry lives at vllm_mlx/aliases.json in the engine repo (audio lives at vllm_mlx/audio/aliases.json). The table on this page mirrors what rapid-mlx models prints.

Naming convention

Every text alias follows the same template:

<family>-<version>-<params>-<modality?>-<technique?>-<quant>
SegmentMeaningExamples
familyModel familyqwen, gemma, llama, mistral, deepseek, phi, hermes
versionMajor version3.5, 3.6, -4, -r1, -v4-flash
paramsParameter count (MoE includes active)4b, 12b, 27b, 35b-a3b (35B total / 3B active)
modality (optional)Non-text variants-vl (vision), -coder (code)
technique (optional)Training-time modifier-qat (QAT), -distill, -thinking, -assistant (MTP assistant drafter)
quant (mandatory)Quantization tier-4bit, -6bit, -8bit, -mxfp4, -mxfp4-q8, -qat-4bit

The quant suffix is mandatory on every alias — qwen3.5-4b-4bit, not qwen3.5-4b. This mirrors the LM Studio …-MLX-4bit HuggingFace convention. -qat is a technique suffix, not a quant — it stacks before the quant, so a QAT-trained Gemma 4 12B at 4-bit is gemma-4-12b-qat-4bit.

Registry schema

Every text entry in aliases.json is a closed dict with a fixed key set. Adding an unknown key (e.g. architecture, family, quantization, or notes) will fail alias loading — copy the field set from the closest existing neighbour when adding a new alias.

KeyTypeMeaning
hf_pathstringHuggingFace repo id in org/name form. The only mandatory field.
tool_call_parserstringWhich tool-call parser to auto-select at serve time (hermes, qwen3_coder, gemma4, harmony, deepseek, …).
reasoning_parserstringWhich reasoning parser to auto-select (qwen3, gemma4, deepseek_r1, gpt_oss, harmony, vibethinker, glm4, minimax, ui_tars).
is_hybridboolHybrid-attention (Mamba / linear-attention / DiffusionGemma / Granite 4 H) — disables spec-decode / suffix.
is_hybrid_explicitboolWhether the hybrid flag was set intentionally (as opposed to defaulted).
is_moeboolMixture-of-experts model.
supports_spec_decodeboolWhether spec-decode routes are eligible.
pflash_tierstringPFlash long-prompt compression tier — verified, unknown, avoid, neutral, n/a.

What rapid-mlx models prints

The rapid-mlx models table has 7 columns. Column semantics match the schema fields above:

ColumnMeaning
AliasThe short name you pass to serve / chat / pull.
ToolsTool-call parser that will be auto-selected (or ).
ReasoningReasoning parser that will be auto-selected (or ).
Spec-Decode when spec-decode is eligible; ✗ hybrid when a hybrid architecture rules it out; when the alias opts out.
Suffix TierSuffix-decoding recommendation: verified, neutral, avoid, unknown, n/a.
DFlashDFlash drafter registry status ( when no registered drafter for this alias).
DDTreeDDTree runtime status ( when not enabled).

Note: the individual spec-decode knobs (Spec-Decode / Suffix Tier / DFlash / DDTree) have since been consolidated under a single --speculative-config option. The exact column layout of rapid-mlx models may differ from the description above; run it locally for the authoritative view.

Run rapid-mlx models locally for the live view — the browse-friendly UI at models.rapidmlx.com renders the same data as a filterable grid.

Text families in 0.12.18

175 text aliases across the following families (count from a byte-exact read of vllm_mlx/aliases.json):

Family# aliasesNotes
Qwen 3 (3.5 · 3.6 · 3.8 · Coder · Coder-Next 80B · VL · legacy)54Anchor family — includes MoE 35B-A3B / 122B, the Coder-Next 80B (qwen3-coder-next-80b-4bit / -8bit, HF-pull), hero coder alias, VL variants, plus the 3.8-27B pair (qwen3.8-27b-4bit / -mixed-3.5bpw).
Gemma 4 (dense · e2b / e4b · qat · assistant · optiq · DiffusionGemma · EmbeddingGemma)31Includes the DiffusionGemma and EmbeddingGemma aliases.
GPT-OSS (20B · 120B · MXFP4-Q4/Q8 · safeguard)10Harmony-native tool + reasoning parsers.
UI-TARS9ByteDance vision-grounded UI automation agent. Custom ui_tars tool parser.
Tmax (9B / 27B, 4/6/8-bit + bf16)7Hybrid Gated-DeltaNet agent model (Ai2). First-mover MLX quants.
Mistral (24B · Small-4 119B · Devstral)6Devstral is the code variant; Small-4 119B is the large MoE SKU.
DeepSeek (R1-Distill · V4-Flash 2/4/8bit + 0731 MXFP4 · Coder-V2-lite)7V4-Flash is single-node-large; R1-Distill covers the 32B slot.
Gemma 3 (1B–27B · qat)6Kept for continuity; new work targets Gemma 4.
Phi 4 (mini + mini-flash 4bit variants)4
Llama (3.1 · 3.2)4
Qwopus3Community fork lineage.
GLM (4.5-Air · 4.7 · 5.2 REAP)3GLM-5.2 is a REAP-50 research SKU.
Liquid (LFM2 24B-A2B · LFM2.5 8B-A1B / 2.6B / 1.2B)4LiquidAI on-device family, lfm parser. HF-pull — see the Liquid page.
MiniMax (M2.5 / M2.7 MXFP4)2Uses the minimax parser pair.
Granite 4 (H-Micro / Tiny)2Hybrid architecture; spec-decode disabled.
VibeThinker (1.5B / 3B)2Low-RAM reasoner — 4-bit fits a 4 GB Mac.
Holo32Hybrid MoE 35B-A3B; operator-owned deployment target.
Hermes 3 · 42
Hunyuan 3 (Hy3)1Tencent 295B MoE preview, Ultra-only. hy_v3 parser; HF-pull — see the Hunyuan page.
Bonsai (Ternary 1.7B · 8B · 27B)32-bit ternary Qwen 3.5-class reasoners. Flagship bonsai-27b-2bit fits a 16 GB Mac (7.9 GB) and is R2-mirrored — see the Bonsai section.
MiniCPM (MiniCPM5 1B · optiq)2OpenBMB on-device family; minicpm parser.
Muse (Glimmer 30B)24-bit + bf16; custom muse tool + reasoning parser.
Ling (3.0 Tiny)1InclusionAI MoE; R2-mirrored. glm47 tool parser.
Others (Kimi K2.6 · Qwen 2.5 · Nemotron 3 / 3.5-Lightning · SmolLM3 · Nanbeige 4 · North-Mini-Code)8One alias each, except Nemotron (3-Nano + 3.5-Lightning) and North-Mini-Code (4-bit + bf16).

Audio registry

Audio aliases live in a separate registry — start them the same way: rapid-mlx serve whisper-large-v3. Registered in 0.12.18:

KindAliases
TTS · kokorokokoro, kokoro-82m, kokoro-82m-bf16, kokoro-82m-4bit, kokoro-82m-8bit, kokoro-4bit, kokoro-8bit
TTS · vibevoice / voxcpm / chatterbox / diavibevoice, vibevoice-realtime, voxcpm, chatterbox, chatterbox-4bit, dia
TTS · qwen3-ttsqwen3-tts, qwen3-tts-4bit, qwen3-tts-6bit, qwen3-tts-clone, qwen3-tts-customvoice, qwen3-tts-voicedesign, qwen3-tts-voicedesign-4bit, qwen3-tts-voicedesign-8bit
TTS · cloning (indextts / f5)f5-tts-zh, indextts, indextts-1.5
STT · whisperwhisper, whisper-1, whisper-large-v3, whisper-large-v3-turbo, whisper-medium, whisper-small, whisper-base, whisper-tiny
STT · parakeetparakeet, parakeet-tdt-0.6b, parakeet-tdt-0.6b-v2, parakeet-v3, parakeet-tdt-0.6b-v3
STT · sensevoicesensevoice, sensevoice-small
STT · qwen3-asrqwen3-asr, qwen3-asr-0.6b, qwen3-asr-1.7b
Forced alignment · qwen3-alignerqwen3-aligner, qwen3-forced-aligner

Audio requires the [audio] extra — pip install "rapid-mlx[audio]==0.12.18". See the extras page for the full dep list.

Video generation registry

Video aliases live in vllm_mlx/aliases.json alongside the text aliases, tagged modality: video-gen — which is why the 175 text count above and the file's raw key count differ. They serve through the asynchronous /v1/videos job API, not /v1/chat/completions. Registered in 0.12.18:

LineAliases
Wan 2.2wan2.2-i2v-a14b-q8, wan2.2-t2v-a14b-bf16, wan2.2-ti2v-5b-bf16, wan2.2-ti2v-5b-q8
CogVideoX-Funcogvideox-fun-5b-bf16, cogvideox-fun-5b-q4, cogvideox-fun-5b-q8
LTX-2.3ltx-2.3-mlx-q4
LTX-2.5ltx-2.5-mlx-q8

Video requires the [video] extra plus ffmpeg and Python 3.11+ — pip install "rapid-mlx[video]==0.12.18". See the video family page.

The (unmapped) column

rapid-mlx ls (aka models --cached) shows every HF repo already on disk. Repos that have no matching alias in aliases.json render as (unmapped) in the Alias column — you can still serve them by their full org/name HF path (e.g. rapid-mlx serve mlx-community/Holo3-35B-A3B-8bit).

Next steps