DeepSeek V4.1 Flash vs Qwen3-VL-30B-A3B-Instruct

DeepSeek V4.1 Flash

A large multimodal agent model with a 1M-token context, documented tool calls, JSON Schema output and a reasoning budget, under MIT. It needs about 612 GB of memory across several GPUs in vLLM or SGLang. The main risks are prompts built from DeepSeek's code rather than a template, no published safety results and no answers in its repositories.

Qwen3-VL-30B-A3B-Instruct

A mid-sized vision-language cousin for screenshots, documents, charts and video that runs fast for its size, with about 3B active parameters. It needs about 23 GB for Alibaba's Q4_K_M GGUF with its vision file, loads in every major runtime under plain Apache 2.0, and three providers host it. With no published safety work, add guardrails before it drives a GUI.

DeepSeek V4.1 Flash

1. DeepSeek's reference run is its inference folder, convert.py with --model-parallel 8 --expert-dtype fp4, then torchrun --nproc-per-node 8 generate.py --ckpt-path "$SAVE_PATH" --config config.json --interactive
2. In vLLM 0.32.0 the parsers are named deepseek_v41 for both --tool-call-parser and --reasoning-parser. That's vLLM's naming, DeepSeek doesn't document a vLLM command
3. Build prompts with encoding.py or pip install deepseek-recipe, not the September repository's chat_template.jinja, which an open pull request says trails the 16 October encoding
4. Tool calls come back in a DSML calls block with a space before calls, invoke and parameter, unlike V4
5. Set reasoning_effort to 1 to 100, or low, high or max (50, 75, 100), in thinking mode only. Default is 75
6. Sample at temperature 1.0 and top_p 1.0 and allow 256K or more output tokens, as the card recommends
7. Without the hardware, call DeepSeek's API as deepseek-flash, which serves this exact update

Qwen3-VL-30B-A3B-Instruct

1. Start it with llama-server -hf Qwen/Qwen3-VL-30B-A3B-Instruct-GGUF:Q4_K_M --jinja -c 32768, or ollama run qwen3-vl:30b-a3b-instruct
2. In Ollama the plain qwen3-vl:30b tag is the Thinking edition, so name 30b-a3b-instruct
3. With the GGUF, load the mmproj vision file too, or images won't be read
4. Parse tool calls as JSON inside <tool_call> tags, or use SGLang's --tool-call-parser qwen
5. Use the shipped sampling, temperature 0.7, top_p 0.8 and top_k 20, from generation_config.json
6. Validate JSON output yourself, since structured output isn't documented
7. vLLM needs 0.11.0 or later, Transformers 4.57.0 or later and llama.cpp b6907 or later