occamy-1.0 vs Qwen3.5-122B-A10B

occamy-1.0

A Qwen3.6-35B-A3B fine-tuned for long agent work with tools, files and code, under plain Apache 2.0, and Accio's 4-bit builds run in about 23 GB. The main gaps are no safety results, agent guidance or security contact, and every benchmark figure is Accio's own.

Qwen3.5-122B-A10B

A strong mid-size cousin for coding, agent steps and reading images or video, with thinking you can switch off per request. Alibaba's 4-bit build needs about 95 GB, which fits one 128 GB unified-memory machine, and several providers host it. The main risk is silence, with no safety results, agent guidance or answers from Alibaba.

occamy-1.0

1. Serve with --reasoning-parser qwen3 --enable-auto-tool-choice --tool-call-parser qwen3_coder on vLLM 0.19.0 or SGLang 0.5.10 and later
2. BF16 is 70.2 GB, so on a 128 GB machine run it without the card's 8-way tensor parallelism
3. Keep the whole assistant message, reasoning and tool calls included, in later turns, as the card says
4. Sampling per the card is temperature 1.0, top_p 0.95, top_k 20 and presence_penalty 1.5
5. With the GGUFs, NFC-normalise input text and read TOKENIZER.md, since the files store qwen2 pre-tokenizer metadata
6. The base's MTP layer isn't in this checkpoint, Accio-Lab/occamy-1.0-MTP is an experimental separate head

Qwen3.5-122B-A10B

1. Quickest start on a 128 GB machine is ollama run qwen3.5:122b (Q4_K_M, 81 GB of files, about 97 GB with context)
2. Alibaba's 4-bit build serves with vllm serve Qwen/Qwen3.5-122B-A10B-GPTQ-Int4 --max-model-len 262144 --reasoning-parser qwen3 --enable-auto-tool-choice --tool-call-parser qwen3_coder --quantization moe_wna16, though that kernel isn't documented for AMD gfx1151
3. Use the qwen3_coder tool parser and the qwen3 reasoning parser, as the card's commands do
4. Turn thinking off with "chat_template_kwargs": {"enable_thinking": false} in the request body, since it's on by default and /think and /nothink aren't supported
5. Sampling per the card is temperature 0.6, top_p 0.95 and top_k 20 when thinking, and temperature 0.7, top_p 0.8, top_k 20 and presence_penalty 1.5 without
6. Keep the context at 128K or more if you rely on thinking, and turn on YaRN only past 262,144 tokens
7. Add --language-model-only in vLLM to skip the vision encoder and free memory for the KV cache when you only need text
8. Validate any JSON it returns, since structured output isn't documented