Independently reproducedpublished Reproduction

Sparrow Chat 3B on 8 GB M1 MacBook Air

Community reproduction confirms the low-memory configuration.

Sparrow Chat 3B @alice Jul 19, 2026

Manifest JSON Reproduce this Run

Speed

18.9 tok/s

Time to first token

651 ms

Peak memory

3.5 GB

Task score

—

Useful

177

6 reproductions

Reproduction command

Use the exact revision and inspect the manifest before execution.

ollama run local-ai/sparrow-chat-3b:q4_k_m

Environment and configuration

The comparison context that makes the result interpretable.

Hardware

MachineApple M1 · 8 GB

CPUApple M1

AcceleratorApple integrated GPU

VRAM8 GB

System memory8 GB

Operating systemmacOS 16

Architecturearm64

Compute APIMetal

Configuration

RuntimeOllama 0.12.1

Model revision5dc3ef1

QuantizationQ4_K_M

Context length2048

Temperature0.7

Seed71

Batch size1

Portable Run manifest

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

{
  "schemaVersion": "1.0",
  "id": "run_sparrow_m1",
  "createdAt": "2026-07-19T06:30:00.000Z",
  "model": {
    "provider": "huggingface",
    "repository": "local-ai/sparrow-chat-3b",
    "revision": "5dc3ef1",
    "quantization": "Q4_K_M",
    "licenseSnapshot": "Apache-2.0"
  },
  "runtime": {
    "name": "Ollama",
    "version": "0.12.1",
    "command": "ollama run local-ai/sparrow-chat-3b:q4_k_m"
  },
  "hardware": {
    "os": "macOS 16",
    "architecture": "arm64",
    "cpu": "Apple M1",
    "systemMemoryMb": 8192,
    "accelerators": [
      {
        "vendor": "Apple",
        "model": "Apple integrated GPU",
        "count": 1,
        "memoryMb": 8192,
        "computeApi": "metal"
      }
    ]
  },
  "task": {
    "type": "chat"
  },
  "configuration": {
    "contextLength": 2048,
    "maxGeneratedTokens": 256,
    "temperature": 0.7,
    "seed": 71,
    "batchSize": 1,
    "quantization": "Q4_K_M"
  },
  "metrics": {
    "timeToFirstTokenMs": 651,
    "tokensPerSecond": 18.9,
    "peakMemoryMb": 3584,
    "modelLoadTimeMs": 1840
  },
  "evidence": {
    "machineCaptured": true,
    "redactionsApplied": [],
    "artifactUrls": []
  }
}

Notes and interpretation

Confirmed with a reduced context window to preserve memory headroom.

Other Runs for Sparrow Chat 3B

Machine-captured2 months ago

Sparrow Chat 3B on a CPU-only laptop

by @mateo · llama.cpp b5881

Speed

12.6 tok/s

TTFT

901 ms

Peak

3.4 GB

Useful

210

Environment and configuration

The comparison context that makes the result interpretable.

Hardware

MachineApple M1 · 8 GB

CPUApple M1

AcceleratorApple integrated GPU

VRAM8 GB

System memory8 GB

Operating systemmacOS 16

Architecturearm64

Compute APIMetal

Configuration

RuntimeOllama 0.12.1

Model revision5dc3ef1

QuantizationQ4_K_M

Context length2048

Temperature0.7

Seed71

Batch size1

{ "schemaVersion": "1.0", "id": "run_sparrow_m1", "createdAt": "2026-07-19T06:30:00.000Z", "model": { "provider": "huggingface", "repository": "local-ai/sparrow-chat-3b", "revision": "5dc3ef1", "quantization": "Q4_K_M", "licenseSnapshot": "Apache-2.0" }, "runtime": { "name": "Ollama", "version": "0.12.1", "command": "ollama run local-ai/sparrow-chat-3b:q4_k_m" }, "hardware": { "os": "macOS 16", "architecture": "arm64", "cpu": "Apple M1", "systemMemoryMb": 8192, "accelerators": [ { "vendor": "Apple", "model": "Apple integrated GPU", "count": 1, "memoryMb": 8192, "computeApi": "metal" } ] }, "task": { "type": "chat" }, "configuration": { "contextLength": 2048, "maxGeneratedTokens": 256, "temperature": 0.7, "seed": 71, "batchSize": 1, "quantization": "Q4_K_M" }, "metrics": { "timeToFirstTokenMs": 651, "tokensPerSecond": 18.9, "peakMemoryMb": 3584, "modelLoadTimeMs": 1840 }, "evidence": { "machineCaptured": true, "redactionsApplied": [], "artifactUrls": [] } }