← All models

BTL-4 (35B-A3B, Q4)

codingApache 2.0toolsreco 10

Bad Theory Labs' agentic build of Ornith 1.0 35B, tuned for tool-heavy coding and long stretches of unattended work. Native reasoning, native tool calling, and a 256K context window that stays light on memory even when it's full. Capable-tier on-device option for machines with 32GB+ memory — see BTL-4 Compact for a smaller build.

At a glance

Parameters
35B
Quantization
Q4_K_M
Context window
262K tokens
Approx. size
21.4 GB
Engines
llama.cpp
License
Apache-2.0 (open)
Version
v1.0.1 , released 2026-08-15
Family
qwen

Tuning

Defaults a gezel client applies out of the box, from real evaluation runs against this exact quantization.

sampling.temperature
0.6
sampling.topP
0.95
sampling.topK
20
sampling.minP
0
sampling.maxTokens
16384
reasoning.thinkingBudget
8192
reasoning.enableThinking
true
profiles
thinking-general, thinking-coding, thinking-precise, instruct, creative

Model behaviors

Client-side behavior modules the catalog enables for this model (reasoning-tag handling, fabrication detection, prompt shaping).

Sources

llama.cpp (GGUF)
bartowski/badtheorylabs_BTL-4-GGUF · badtheorylabs_BTL-4-Q4_K_M.gguf · Q4_K_M
Upstream
https://huggingface.co/badtheorylabs/BTL-4
Manifest
View on GitHub

Tags