Inkling vs Qwen3.5-0.8B

Inkling

Thinking Machines' large multimodal generalist, stronger than Inkling-Small on factual recall by its own numbers. It needs about 710 GB of datacentre GPU memory even in NVFP4, so most agents will use a host such as Together AI or Tinker. The main risks are a use policy that binds on download and can change, and no outside safety results.

Qwen3.5-0.8B

A tiny vision-language cousin for prototypes, fine-tuning and light local steps such as routing, tagging or reading an image, running in under 2 GB in every major runtime under plain Apache 2.0. Keep thinking off, since the card warns it loops. The main risk is neglect, with no hosted option, safety results or answers from Alibaba since March.

Inkling

1. Unless you have a multi-GPU Blackwell or H200 node, call it through Together AI, Fireworks, DeepInfra or Tinker rather than starting it
2. On 8 B200 GPUs, SGLang's recipe is sglang serve --trust-remote-code --model-path thinkingmachines/Inkling-NVFP4 --tp 8 --quantization modelopt_fp4 --enable-multimodal --reasoning-parser inkling --tool-call-parser inkling
3. Use the inkling tool parser, not Hermes, and the inkling reasoning parser
4. Pass reasoning_effort as none, minimal, low, medium, high (the default) or max, or a number from 0 to 0.99. The repository's template has no xhigh preset and raises an error for it, though Thinking Machines' docs list xhigh
5. Tool call arguments must reach the template as an object, not a JSON string, or it raises an error
6. Audio should be 16 kHz WAV, ideally under 20 minutes, and images 40 to 4096 px a side, per the card
7. Read the Model Acceptable Use Policy before deploying, it binds on download and bars some uses that are legal

Qwen3.5-0.8B

1. Smallest start is ollama run qwen3.5:0.8b, or qwen3.5:0.8b-q8_0 at 1.0 GB
2. Full precision vllm serve Qwen/Qwen3.5-0.8B --port 8000 --max-model-len 262144 --enable-auto-tool-choice --tool-call-parser qwen3_coder
3. Thinking is off by default; if you turn it on with "chat_template_kwargs": {"enable_thinking": true}, stream and cut off runaway output, since the card warns of thinking loops, and add --reasoning-parser qwen3 in vLLM or SGLang to split the reasoning out
4. There's no generation_config.json, so pass the card's sampling yourself, for example temperature=1.0, top_p=1.0, top_k=20, presence_penalty=2.0 for text without thinking
5. Use it for narrow steps or as a fine-tuning base, not as the main agent, as the card itself suggests
6. Validate any JSON it returns, since structured output isn't documented