Apodex-1.1-mini vs Qwen3.5-397B-A17B

Apodex-1.1-mini

An agent cousin fine-tuned from Qwen3.5-35B-A3B for long tasks with files, code and tools, with parallel tool calling and about 3B active parameters. It needs 32 GB for Apodex's GGUF, NVFP4 or GPTQ-Int4 builds or about 86 GB in BF16, and Apodex's API serves it too. The main risks are no safety results and no replies from the publisher.

Qwen3.5-397B-A17B

Alibaba's biggest open Qwen3.5, a cousin for hard coding, agent and vision work. Even at 4 bits it needs about 270 GB, so most agents will call it through Alibaba's Model Studio or one of the many providers that host it. The main risk is silence, with no safety results, no agent guidance and template reports without a reply.

Apodex-1.1-mini

1. Serve it with vllm serve apodex/Apodex-1.1-mini --enable-auto-tool-choice --tool-call-parser qwen3_coder --reasoning-parser qwen3, or the same parsers in SGLang, so tool calls and reasoning come back parsed
2. For llama.cpp, llama-server -hf apodex/Apodex-1.1-mini-GGUF:Q4_K_M --jinja --reasoning-format deepseek, and the card passes the repository's chat_template.jinja with --chat-template-file
3. Pass tools through the API's tools parameter rather than in the system prompt, as the card asks
4. The card recommends temperature 1.0, top_p 0.95, repetition_penalty 1.05 and max_tokens up to 32,768
5. Send your own system prompt, or set identity_mode, since the template otherwise adds an Apodex identity block
6. The template keeps all earlier reasoning in the context unless preserve_thinking is false, which grows long agent sessions
7. Thinking is on by default, send chat_template_kwargs with enable_thinking false for short tasks, a switch the card doesn't mention
8. The card's Python example has a broken line where it runs the tool (available_toolsname), so fix it to available_tools[name](**args) before running
9. Validate JSON output yourself, since structured output isn't documented

Qwen3.5-397B-A17B

1. The card's server command is vllm serve Qwen/Qwen3.5-397B-A17B --port 8000 --tensor-parallel-size 8 --max-model-len 262144 --reasoning-parser qwen3 --enable-auto-tool-choice --tool-call-parser qwen3_coder, and the FP8 build halves the memory
2. It doesn't fit our 128 GB reference machine at any precision, so call a hosted endpoint instead (Model Studio's Qwen3.5-Plus, or a provider serving Qwen/Qwen3.5-397B-A17B)
3. On Model Studio pass "enable_thinking": false directly, while vLLM and SGLang take "chat_template_kwargs": {"enable_thinking": false}
4. 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
5. For multi-token prediction in vLLM add --speculative-config '{"method":"qwen3_next_mtp","num_speculative_tokens":2}'
6. Keep the context at 128K or more if you rely on thinking, and turn on YaRN only past 262,144 tokens
7. Validate any JSON it returns, since structured output isn't documented