Independently reproducedpublished Reproduction

Aurora Reason 8B on M2 Max

Equivalent configuration reproduced on Apple Silicon.

Aurora Reason 8B @alice Jul 16, 2026

Manifest JSON Reproduce this Run

Speed

29.7 tok/s

Time to first token

390 ms

Peak memory

8.6 GB

Reasoning suite

67.9

Useful

3 reproductions

Reproduction command

Use the exact revision and inspect the manifest before execution.

modelverse reproduce run_aurora_4090_reasoning --runtime mlx

Environment and configuration

The comparison context that makes the result interpretable.

Hardware

MachineApple M2 Max · 32 GB

CPUApple M2 Max

AcceleratorApple integrated GPU

VRAM32 GB

System memory32 GB

Operating systemmacOS 16

Architecturearm64

Compute APIMetal

Configuration

RuntimeMLX 0.29.2

Model revisiona887c31

QuantizationQ5_K_M

Context length16384

Temperature0

Seed42

Batch size1

Portable Run manifest

Version 1.0 of the Apache-2.0 licensed Modelverse Run schema.

{
  "schemaVersion": "1.0",
  "id": "run_aurora_m2max",
  "createdAt": "2026-07-16T09:20:00.000Z",
  "model": {
    "provider": "huggingface",
    "repository": "open-lab/aurora-reason-8b",
    "revision": "a887c31",
    "quantization": "Q5_K_M",
    "licenseSnapshot": "Apache-2.0"
  },
  "runtime": {
    "name": "MLX",
    "version": "0.29.2",
    "command": "modelverse reproduce run_aurora_4090_reasoning --runtime mlx"
  },
  "hardware": {
    "os": "macOS 16",
    "architecture": "arm64",
    "cpu": "Apple M2 Max",
    "systemMemoryMb": 32768,
    "accelerators": [
      {
        "vendor": "Apple",
        "model": "Apple integrated GPU",
        "count": 1,
        "memoryMb": 32768,
        "computeApi": "metal"
      }
    ]
  },
  "task": {
    "type": "reasoning",
    "evaluationSuite": "modelverse-reasoning-v1",
    "datasetRevision": "0f71fd2"
  },
  "configuration": {
    "contextLength": 16384,
    "maxGeneratedTokens": 2048,
    "temperature": 0,
    "seed": 42,
    "batchSize": 1,
    "quantization": "Q5_K_M"
  },
  "metrics": {
    "timeToFirstTokenMs": 390,
    "tokensPerSecond": 29.7,
    "peakMemoryMb": 8806.4,
    "modelLoadTimeMs": 3070,
    "scores": {
      "Reasoning suite": 67.9
    }
  },
  "evidence": {
    "machineCaptured": true,
    "redactionsApplied": [],
    "artifactUrls": []
  }
}

Notes and interpretation

Score falls inside the declared reproduction tolerance.

Evaluation suite

modelverse-reasoning-v1

dataset revision 0f71fd2

Other Runs for Aurora Reason 8B

Standardized evaluation2 months ago

Aurora Reason 8B standardized reasoning evaluation

by @chen · llama.cpp b5881

Speed

56.9 tok/s

TTFT

226 ms

Peak

8.3 GB

Useful

172

Environment and configuration

The comparison context that makes the result interpretable.

Hardware

MachineApple M2 Max · 32 GB

CPUApple M2 Max

AcceleratorApple integrated GPU

VRAM32 GB

System memory32 GB

Operating systemmacOS 16

Architecturearm64

Compute APIMetal

Configuration

RuntimeMLX 0.29.2

Model revisiona887c31

QuantizationQ5_K_M

Context length16384

Temperature0

Seed42

Batch size1

{ "schemaVersion": "1.0", "id": "run_aurora_m2max", "createdAt": "2026-07-16T09:20:00.000Z", "model": { "provider": "huggingface", "repository": "open-lab/aurora-reason-8b", "revision": "a887c31", "quantization": "Q5_K_M", "licenseSnapshot": "Apache-2.0" }, "runtime": { "name": "MLX", "version": "0.29.2", "command": "modelverse reproduce run_aurora_4090_reasoning --runtime mlx" }, "hardware": { "os": "macOS 16", "architecture": "arm64", "cpu": "Apple M2 Max", "systemMemoryMb": 32768, "accelerators": [ { "vendor": "Apple", "model": "Apple integrated GPU", "count": 1, "memoryMb": 32768, "computeApi": "metal" } ] }, "task": { "type": "reasoning", "evaluationSuite": "modelverse-reasoning-v1", "datasetRevision": "0f71fd2" }, "configuration": { "contextLength": 16384, "maxGeneratedTokens": 2048, "temperature": 0, "seed": 42, "batchSize": 1, "quantization": "Q5_K_M" }, "metrics": { "timeToFirstTokenMs": 390, "tokensPerSecond": 29.7, "peakMemoryMb": 8806.4, "modelLoadTimeMs": 3070, "scores": { "Reasoning suite": 67.9 } }, "evidence": { "machineCaptured": true, "redactionsApplied": [], "artifactUrls": [] } }