LBS-27B vs Qwen3.8-Flash-Next

LBS-27B

A 27B Qwen3.8 fine-tune for GUI grounding that zooms into screenshots over a few turns before naming a point, and inclusionAI says it also helps with fine detail in large images. It needs about 96 GB at BF16 and a client that runs its crop actions. Every result so far is inclusionAI's own, with no safety results or injection guidance.

Qwen3.8-Flash-Next

A large sparse cousin with 6B active parameters for agentic coding and multimodal work, previewing Qwen's next architecture. It needs a 128 GB Mac for Ollama's MLX build or about 144 GB for the 4-bit GGUF, and Qwen Cloud serves it. The main risk is a licence that needs a separate deal for hosted-model and coding or office assistant businesses.

LBS-27B

1. Serve with vllm serve inclusionAI/LBS-27B --tensor-parallel-size 4 --max-model-len 65536, which is what the card's wrapper script runs; keep the context at 32,768 or more for multi-turn images
2. Build the system prompt with build_system_prompt from the LBS branch's zoomin_grounding_protocol.py, send the screenshot after image_path: <name>, and don't pass tools; the model answers in plain text
3. Parse <action>crop_image(image_path=..., x1=..., y1=..., x2=..., y2=...)</action>, cut the crop yourself and return it with its mapping to the original image; allow at most two crops per turn and three per episode
4. Read the final answer as [x.xx,y.yy] on a 0 to 1000 scale of the original image, and use your own executor to click
5. Pass reasoning_effort as xhigh, medium or low only, since the template raises an error for high, and don't count on thinking_budget, which the template ignores
6. Don't turn on MTP speculative decoding; the MTP weights aren't in the repository though config.json names one MTP layer

Qwen3.8-Flash-Next

1. On a 128 GB Mac start it with ollama run qwen3.8-flash-next:125b-mlx (105 GB, NVFP4), on Linux the smallest Ollama build is qwen3.8-flash-next:125b-a6b-q4_K_M at 120 GB, which won't fit a 107.4 GB GPU share
2. On a server, vllm serve Qwen/Qwen3.8-Flash-Next-FP8 --tensor-parallel-size 4 --max-model-len 262144 --reasoning-parser qwen3 --enable-auto-tool-choice --tool-call-parser qwen3_coder
3. Use the qwen3_coder tool parser and the qwen3 reasoning parser, as Qwen's GitHub README does
4. Thinking is on by default, pass chat_template_kwargs with enable_thinking false to skip it, and read the reasoning from reasoning_content, even though SGLang's cookbook says thinking can't be turned off
5. reasoning_effort takes xhigh (default), medium or low, and any other value makes the template raise an error
6. Send at most one system message and put it first, since the template raises an error otherwise
7. Sampling per the card is temperature 1.0 and top_p 0.95 when thinking, temperature 0.7, top_p 0.8 and presence_penalty 1.5 without
8. vLLM keeps the 51B n-gram table in host memory on 80 GB GPUs by default, per its recipe
9. Validate JSON output yourself or use the runtime's guided decoding, since structured output isn't documented
10. Read the Qwen Community License before serving it to others as an API or inside a coding or office assistant product