Downloads · 30 days
42
3% of all-time downloads
rachpradhan/Qwen3.5-35B-A3B-Turbo-SWE-v0.0.1
Qwen3.5-35B-A3B-Turbo-SWE-v0.0.1 is a text generation model from rachpradhan. Use it when you need the model to write or continue text. The card lists the license as apache-2.0.
A coding-focused fine-tune of Qwen3.5-35B-A3B using SFT + GRPO on a mixture of real-world coding agent trajectories from Codex, Claude Code, and OpenCode sessions.
Downloads · 30 days
42
3% of all-time downloads
All-time downloads
1.6K
Public
Repo size
21.2 GB
Likes
8
Public
Click a slice to open those files.
.gguf21.2 GB · 100%
From the Hugging Face model README
A coding-focused fine-tune of Qwen3.5-35B-A3B using SFT + GRPO on a mixture of real-world coding agent trajectories from Codex, Claude Code, and OpenCode sessions.
Qwen3.5-35B-A3B is a 35B parameter MoE model with only 3B active parameters per token. This fine-tune improves its coding capabilities by training on real-world software engineering trajectories.
| Property | Value |
|---|---|
| Base Model | Qwen3.5-35B-A3B |
| Total Parameters | 35B |
| Active Parameters | ~3B per token |
| Experts | 256 total, 8 routed + 1 shared |
| Attention | Hybrid: 30x Gated DeltaNet + 10x Full Attention |
| Context | 262K native |
| Quantization | Q4_K_M (4.88 BPW) |
| Metric | Value |
|---|---|
| Prompt eval | 139 tok/s (M3 Ultra) |
| Decode | 70.7 tok/s (M3 Ultra) |
| Model size | 20GB (Q4_K_M) |
| VRAM usage | ~21GB |
llama-server -m Qwen3.5-35B-A3B-Turbo-SWE-v0.0.1-Q4_K_M.gguf --port 8080
echo 'FROM Qwen3.5-35B-A3B-Turbo-SWE-v0.0.1-Q4_K_M.gguf' > Modelfile
ollama create turbo-swe -f Modelfile
ollama run turbo-swe
from openai import OpenAI
client = OpenAI(base_url="http://localhost:8080/v1", api_key="none")
response = client.chat.completions.create(
model="turbo-swe",
messages=[{"role": "user", "content": "Write a prime checker in Python"}],
)
print(response.choices[0].message.content)
Apache 2.0 (same as base model)