badtheorylabs / BTL-3-Compact

Overview

BTL-3 Compact is the portable native edition of BTL-3. It packages the complete text model—including the compressed decoder, vocabulary matrices, output correction, and behavior repair—into a single 8,392,369,600-byte GGUF .

The runtime consumes the packed representation directly. It does not download, load, or reconstruct the original BF16 checkpoint.

Highlights

• Complete Qwen3.6-27B-derived text model in 8.39 GB decimal / 7.82 GiB .
• Full 64-layer mixed AVQ2/UniSVQ representation with targeted precision islands and compact behavioral corrections.
• 92.2% overall conditional retention on a fresh 100-turn tool-contract gate authored after compression choices were frozen.
• 83/100 absolute compact accuracy against a full-precision teacher that scored 90/100 on the same sealed gate.
• Exact native CUDA and Metal execution with no persistent dense reconstruction.
• OpenAI-compatible local server plus LM Studio and Ollama CLI integrations.
• Verified macOS arm64 package and a clearly separated DGX Spark CUDA preview.

Model specification

Usable context depends on device memory, KV cache, and runtime workspace. Start with 4K on a 16 GB Mac, 16K on a 12–16 GB GPU, or 32K on a 24 GB GPU, then increase the window after measuring headroom.

Measured retention

The fresh gate contains 100 turns balanced across single, parallel, sequential, parallel-multiple, and abstention behavior. It is an internal contract-retention evaluation, not a public frontier benchmark.

The complete protocol and category report are included in evidence/compact-validation.md .

Native performance

The RTX result is the mean of three native runs with a 512-token prompt and 128 generated tokens. The M2 measurement is a compatibility smoke on the entry-level 16 GB system, not a performance projection for newer Apple hardware.

Runtime support

Stock Ollama and the stock LM Studio GGUF engine do not decode AVQ2 directly. The included integrations start or connect to BTL's native runner while preserving the familiar client interface.

Quickstart

Install the verified macOS package

python3 tools/install_consumer_bundle.py \
  --runtime runtimes/supported/BTL-3-Compact-macos-arm64 \
  --model model/BTL-3-Compact-AVQ2.gguf

Start the server directly

BTL3_MODEL="$PWD/model/BTL-3-Compact-AVQ2.gguf" \
BTL3_CTX_SIZE=4096 \
  runtimes/supported/BTL-3-Compact-macos-arm64/bin/btl3-server

Call the OpenAI-compatible API

curl http://127.0.0.1:8080/v1/chat/completions \
  -H 'Content-Type: application/json' \
  -d '{
    "model": "BTL-3",
    "messages": [
      {
        "role": "user",
        "content": "Write a retrying fetch helper and include tests."
      }
    ],
    "stream": true
  }'

Representation

The release combines:

• packed AVQ2 decoder tensors;
• affine INT4 tensors and two measured INT4 demotions;
• selected higher-precision islands;
• packed embedding and output matrices;
• a rank-32 output correction;
• a compact behavior adapter.

All 2,416 tensor payloads were byte-verified during export. Native CUDA parity was checked against the reference representation, and the exact exported GGUF subsequently completed autoregressive generation on CUDA and Metal.

Artifact integrity

Verify the complete release directory:

shasum -a 256 -c SHA256SUMS

RELEASE_MANIFEST.json records the exact model identity and separates verified runtimes from preview packages.

Intended use

• local coding and debugging;
• private repository and terminal agents;
• structured single, sequential, and parallel tool use;
• offline or self-hosted OpenAI-compatible inference;
• consumer and workstation deployments where the BF16 model is impractical.

Operational guidance

Run generated code and tool calls in a sandbox. Require explicit confirmation before destructive, privileged, financial, or otherwise high-impact actions.

License and citation

The model artifact is Apache-2.0. The included runtime is MIT-licensed; see the packaged licenses and third-party notices.

@software{btl3_compact_2026,
  title  = {BTL-3 Compact: A Native 8.39 GB Agentic Coding Model},
  author = {Bad Theory Labs},
  year   = {2026},
  url    = {https://huggingface.co/badtheorylabs/BTL-3-Compact}
}

For questions and release updates, visit Bad Theory Labs or join the community Discord .