Text Generation
MLX
Safetensors
qwen3_5_moe
apple-silicon
quantized
mixed-precision
axquant
axq
development
qwen3.5-moe
MXFP8
mtp
vision
conversational
8-bit precision
Instructions to use AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP with MLX:
# Make sure mlx-lm is installed # pip install --upgrade mlx-lm # Generate text with mlx-lm from mlx_lm import load, generate model, tokenizer = load("AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP") prompt = "Write a story about Einstein" messages = [{"role": "user", "content": prompt}] prompt = tokenizer.apply_chat_template( messages, add_generation_prompt=True ) text = generate(model, tokenizer, prompt=prompt, verbose=True) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Pi
How to use AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP with Pi:
Start the MLX server
# Install MLX LM: uv tool install mlx-lm # Start a local OpenAI-compatible server: mlx_lm.server --model "AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP"
Configure the model in Pi
# Install Pi: npm install -g @earendil-works/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "mlx-lm": { "baseUrl": "http://localhost:8080/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP" } ] } } }Run Pi
# Start Pi in your project directory: pi
- MLX LM
How to use AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP with MLX LM:
Generate or start a chat session
# Install MLX LM uv tool install mlx-lm # Interactive chat REPL mlx_lm.chat --model "AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP"
Run an OpenAI-compatible server
# Install MLX LM uv tool install mlx-lm # Start the server mlx_lm.server --model "AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP" # Calling the OpenAI-compatible server with curl curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP", "messages": [ {"role": "user", "content": "Hello"} ] }' - Hermes Agent
How to use AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP with Hermes Agent:
Start the MLX server
# Install MLX LM: uv tool install mlx-lm # Start a local OpenAI-compatible server: mlx_lm.server --model "AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP"
Configure Hermes
# Install Hermes: curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash hermes setup # Point Hermes at the local server: hermes config set model.provider custom hermes config set model.base_url http://127.0.0.1:8080/v1 hermes config set model.default AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP with OpenClaw:
Start the MLX server
# Install MLX LM: uv tool install mlx-lm # Start a local OpenAI-compatible server: mlx_lm.server --model "AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP"
Configure OpenClaw
# Install OpenClaw: npm install -g openclaw@latest # Register the local server and set it as the default model: openclaw onboard --non-interactive --mode local \ --auth-choice custom-api-key \ --custom-base-url http://127.0.0.1:8080/v1 \ --custom-model-id "AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP" \ --custom-provider-id mlx-lm \ --custom-compatibility openai \ --custom-text-input \ --accept-risk \ --skip-health
Run OpenClaw
openclaw agent --local --agent main --message "Hello from Hugging Face"
Correct evidenced MXFP8 container mode and bind runtime format audit
Browse files- README.md +13 -0
- axquant_manifest.json +10 -5
- config.json +2 -2
- runtime_audit.json +82 -0
README.md
CHANGED
|
@@ -19,6 +19,19 @@ tags:
|
|
| 19 |
- vision
|
| 20 |
---
|
| 21 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 22 |
# AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP — 8.46 BPW measured main
|
| 23 |
|
| 24 |
An **AXQuant (AXQ)** mixed-precision MLX checkpoint for Apple Silicon, converted directly from
|
|
|
|
| 19 |
- vision
|
| 20 |
---
|
| 21 |
|
| 22 |
+
## Runtime format audit (2026-10-06)
|
| 23 |
+
|
| 24 |
+
The quantization container mode was corrected from `affine` to `mxfp8`
|
| 25 |
+
in both config blocks after inspecting every quantized module header.
|
| 26 |
+
Weight bytes and per-module precision assignments are unchanged.
|
| 27 |
+
This family has no n-gram tensors; no n-gram file or declaration was added.
|
| 28 |
+
This remote format audit does not grant an oMLX/MTPLX runtime profile or a successful load/generation claim.
|
| 29 |
+
|
| 30 |
+
See [runtime_audit.json](runtime_audit.json) for pinned config/index/header
|
| 31 |
+
bindings, architecture, physical-format findings, and applied corrections.
|
| 32 |
+
This is development evidence; no quality, MTP exactness, speed, or certification
|
| 33 |
+
claim is added. Historical evidence stays bound to its original revision.
|
| 34 |
+
|
| 35 |
# AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP — 8.46 BPW measured main
|
| 36 |
|
| 37 |
An **AXQuant (AXQ)** mixed-precision MLX checkpoint for Apple Silicon, converted directly from
|
axquant_manifest.json
CHANGED
|
@@ -1,13 +1,13 @@
|
|
| 1 |
{
|
| 2 |
"axquant_version": "1.9.0",
|
| 3 |
"calibration": null,
|
| 4 |
-
"created_at": "2026-10-
|
| 5 |
"effective_bpw": 9.350650324235088,
|
| 6 |
"files": [
|
| 7 |
{
|
| 8 |
"path": "README.md",
|
| 9 |
-
"sha256": "
|
| 10 |
-
"size_bytes":
|
| 11 |
},
|
| 12 |
{
|
| 13 |
"path": "ax_expert_stream.json",
|
|
@@ -51,8 +51,8 @@
|
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"path": "config.json",
|
| 54 |
-
"sha256": "
|
| 55 |
-
"size_bytes":
|
| 56 |
},
|
| 57 |
{
|
| 58 |
"path": "generation_config.json",
|
|
@@ -124,6 +124,11 @@
|
|
| 124 |
"sha256": "d89ef49ce9cd37fbf510158e13c1ef063d9286411c1ec9049932dbe0487143b1",
|
| 125 |
"size_bytes": 1191
|
| 126 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 127 |
{
|
| 128 |
"path": "tokenizer.json",
|
| 129 |
"sha256": "5f9e4d4901a92b997e463c1f46055088b6cca5ca61a6522d1b9f64c4bb81cb42",
|
|
|
|
| 1 |
{
|
| 2 |
"axquant_version": "1.9.0",
|
| 3 |
"calibration": null,
|
| 4 |
+
"created_at": "2026-10-06T21:00:28.724844+00:00",
|
| 5 |
"effective_bpw": 9.350650324235088,
|
| 6 |
"files": [
|
| 7 |
{
|
| 8 |
"path": "README.md",
|
| 9 |
+
"sha256": "6806b2a2c7168c4d1056290bb0c946b4ce553a54c488a3c2c4253c2b0c8b7ece",
|
| 10 |
+
"size_bytes": 9302
|
| 11 |
},
|
| 12 |
{
|
| 13 |
"path": "ax_expert_stream.json",
|
|
|
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"path": "config.json",
|
| 54 |
+
"sha256": "19ba4f1e9ca1e1576bf2b3dd1968ce59facb31ea75f4f37271db9e1b01f5bec3",
|
| 55 |
+
"size_bytes": 138633
|
| 56 |
},
|
| 57 |
{
|
| 58 |
"path": "generation_config.json",
|
|
|
|
| 124 |
"sha256": "d89ef49ce9cd37fbf510158e13c1ef063d9286411c1ec9049932dbe0487143b1",
|
| 125 |
"size_bytes": 1191
|
| 126 |
},
|
| 127 |
+
{
|
| 128 |
+
"path": "runtime_audit.json",
|
| 129 |
+
"sha256": "850e38fa7ec1eaacbd11831f77dec8ef10c9ab160332ece2031043a5ee9d534a",
|
| 130 |
+
"size_bytes": 3954
|
| 131 |
+
},
|
| 132 |
{
|
| 133 |
"path": "tokenizer.json",
|
| 134 |
"sha256": "5f9e4d4901a92b997e463c1f46055088b6cca5ca61a6522d1b9f64c4bb81cb42",
|
config.json
CHANGED
|
@@ -2570,7 +2570,7 @@
|
|
| 2570 |
"group_size": 32,
|
| 2571 |
"mode": "mxfp8"
|
| 2572 |
},
|
| 2573 |
-
"mode": "
|
| 2574 |
},
|
| 2575 |
"quantization_config": {
|
| 2576 |
"bits": 8,
|
|
@@ -5130,7 +5130,7 @@
|
|
| 5130 |
"group_size": 32,
|
| 5131 |
"mode": "mxfp8"
|
| 5132 |
},
|
| 5133 |
-
"mode": "
|
| 5134 |
},
|
| 5135 |
"text_config": {
|
| 5136 |
"attention_bias": false,
|
|
|
|
| 2570 |
"group_size": 32,
|
| 2571 |
"mode": "mxfp8"
|
| 2572 |
},
|
| 2573 |
+
"mode": "mxfp8"
|
| 2574 |
},
|
| 2575 |
"quantization_config": {
|
| 2576 |
"bits": 8,
|
|
|
|
| 5130 |
"group_size": 32,
|
| 5131 |
"mode": "mxfp8"
|
| 5132 |
},
|
| 5133 |
+
"mode": "mxfp8"
|
| 5134 |
},
|
| 5135 |
"text_config": {
|
| 5136 |
"attention_bias": false,
|
runtime_audit.json
ADDED
|
@@ -0,0 +1,82 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"applied_config_corrections": [
|
| 3 |
+
{
|
| 4 |
+
"field": "quantization.mode",
|
| 5 |
+
"new": "mxfp8",
|
| 6 |
+
"old": "affine",
|
| 7 |
+
"physical_evidence_consistent": true
|
| 8 |
+
},
|
| 9 |
+
{
|
| 10 |
+
"field": "quantization_config.mode",
|
| 11 |
+
"new": "mxfp8",
|
| 12 |
+
"old": "affine",
|
| 13 |
+
"physical_evidence_consistent": true
|
| 14 |
+
}
|
| 15 |
+
],
|
| 16 |
+
"current_config_sha256": "19ba4f1e9ca1e1576bf2b3dd1968ce59facb31ea75f4f37271db9e1b01f5bec3",
|
| 17 |
+
"date": "2026-10-06",
|
| 18 |
+
"header_sha256": {
|
| 19 |
+
"model-00001-of-00007.safetensors": "10ff1b6d08c57bf5a77f0d8c2a6a701498ed04284b36d9d56fcba0ebf854f221",
|
| 20 |
+
"model-00002-of-00007.safetensors": "6cd799227928a226b05d9816966dc6ffab93124762f993e96c27328a35bef543",
|
| 21 |
+
"model-00003-of-00007.safetensors": "5fc7dc14699a6bc95a3a3196b1e19883cf7bb2940fc193d3b24326791492ea66",
|
| 22 |
+
"model-00004-of-00007.safetensors": "dba7c367c038632a91f5b8afd3186096c724c257e3b69eba0e22730ff61b5139",
|
| 23 |
+
"model-00005-of-00007.safetensors": "7d7d8ee10cfc768aa1e685ca7c67ecbcf38daf3157db8b298604a0f63cf19392",
|
| 24 |
+
"model-00006-of-00007.safetensors": "5ffbef9f343c1b09483856096c801948a5b72a752a1d18e1ce96c59ce20ee1a0",
|
| 25 |
+
"model-00007-of-00007.safetensors": "1057d271cab088366b759331b14c0c367d182927f826210baed1d47fe40eaad5",
|
| 26 |
+
"mtp.safetensors": "e5f315b744dc3017e055dd1be599554bbaf6edd331447ec0f7f1416accab5196",
|
| 27 |
+
"vision.safetensors": "3901371df301f62a04c09f693bfdab663309708fd6cca251493366e0b616eb1e"
|
| 28 |
+
},
|
| 29 |
+
"input_bindings": {
|
| 30 |
+
"axquant_manifest.json": "4bb1dbc3c21cef5331dbf82603c56a53b4ea7fceef99da0fd3655aec09d86e4d",
|
| 31 |
+
"axquant_mtp_sidecar_manifest.json": "3d0ac7b666aa175f47424480867c5ca921aa2e801aa8cbdba2aa8cc578ece0da",
|
| 32 |
+
"config.json": "8990d524663b982768fb76f10dd10becf53bc8fa630943900e3da61debe86b42",
|
| 33 |
+
"model.safetensors.index.json": "d2b1bd9647b037dd7b7b33e5d5b49027c4f551fb3aa0676868af7b258b27cb4c",
|
| 34 |
+
"mtplx_runtime.json": "f004052d2cf52bd7bab68d30ade118cab28484394b5eb6b2271127a08ee70138"
|
| 35 |
+
},
|
| 36 |
+
"issues": [],
|
| 37 |
+
"model_type": "qwen3_5_moe",
|
| 38 |
+
"mtp_files": [
|
| 39 |
+
"mtp.safetensors"
|
| 40 |
+
],
|
| 41 |
+
"ngram_action": "none; do not invent n-gram data",
|
| 42 |
+
"ngram_files": [],
|
| 43 |
+
"ngram_quantization": [],
|
| 44 |
+
"ngram_table_metadata": null,
|
| 45 |
+
"ngram_tensor_count": 0,
|
| 46 |
+
"quality_certified": false,
|
| 47 |
+
"quantization": {
|
| 48 |
+
"quantization": {
|
| 49 |
+
"container_mode": "affine",
|
| 50 |
+
"per_module_modes": {
|
| 51 |
+
"mxfp8": 511
|
| 52 |
+
},
|
| 53 |
+
"physical_recipe_issues": []
|
| 54 |
+
},
|
| 55 |
+
"quantization_config": {
|
| 56 |
+
"container_mode": "affine",
|
| 57 |
+
"per_module_modes": {
|
| 58 |
+
"mxfp8": 511
|
| 59 |
+
},
|
| 60 |
+
"physical_recipe_issues": []
|
| 61 |
+
}
|
| 62 |
+
},
|
| 63 |
+
"removed_source_evidence": [],
|
| 64 |
+
"repo_id": "AutomatosX/AX-Cyber-Tiel-Coder-35B-A3B-MLX-AXQ-MXFP8-MTP",
|
| 65 |
+
"runtime_arch_id": null,
|
| 66 |
+
"runtime_verified": false,
|
| 67 |
+
"schema_version": "axquant.hub-runtime-audit.v1",
|
| 68 |
+
"scope": "Pinned remote config/index/Safetensors header audit; no runtime load or generation claim.",
|
| 69 |
+
"source_revision": "b68a546e4e82030f0591740b8408d34ac9c07674",
|
| 70 |
+
"status": "no-audited-peer-export-profile",
|
| 71 |
+
"unchanged_weight_sha256": {
|
| 72 |
+
"model-00001-of-00007.safetensors": "2432953bbd6048c30d356c373dbd3f549dcd3fc936c3cd857495e7c389b87251",
|
| 73 |
+
"model-00002-of-00007.safetensors": "001d5e0d8384b95f3b521486999df4edf228bcfc45779cabefdcf89b9e9f9c54",
|
| 74 |
+
"model-00003-of-00007.safetensors": "09481882a8e50669c716adb2b08bc45f429d78df3a03ce9eeb9400aa56b7e746",
|
| 75 |
+
"model-00004-of-00007.safetensors": "bf7c520b0a76410398ebae1aa92e1de507ad7813fbb8a87f7d936831c845266a",
|
| 76 |
+
"model-00005-of-00007.safetensors": "4455111dcd9e064243c6df6726cd115dcffbbdfe722cced6051a9c8845cf8f4c",
|
| 77 |
+
"model-00006-of-00007.safetensors": "8c41a9f7bfe78e5b9a7e1e5ffeea5bf64f0a148eabb8a64c16210de384c9239c",
|
| 78 |
+
"model-00007-of-00007.safetensors": "f2a251a5aa48071096ad1aeebaad9637b6fd641d1e54f3d5f6700c2071e3db0e",
|
| 79 |
+
"mtp.safetensors": "590e87c9c3fbbaa370c8dddbcc22611f00bbc78019d28e3192f180841c030527",
|
| 80 |
+
"vision.safetensors": "56ee429e45e42b276c8a94fb84b5391994d5c701b24880a74b0908f2f1ffd251"
|
| 81 |
+
}
|
| 82 |
+
}
|