Repository
Qwen3.6-35b-vs-DSF4-tool-eval-bench
Tool evaluation benchmarks comparing Qwen3.6-35B vs DeepSeek V4 Flash DSpark
- File
- Description
- Qwen3.6-35B-A3B-UD-Q8_K_XL.html
- Tool evaluation results for the Qwen3.6-35B model
- DeepSeek-V4-Flash-DSpark.html
- Tool evaluation results for the DeepSeek V4 Flash DSpark model
- DeepSeek-V4-vs-Qwen3.6-35B_comparison.html
- Side-by-side comparison of both models across all benchmark dimensions
Specs
- FileDescription
- Qwen3.6-35B-A3B-UD-Q8_K_XL.htmlTool evaluation results for the Qwen3.6-35B model
- DeepSeek-V4-Flash-DSpark.htmlTool evaluation results for the DeepSeek V4 Flash DSpark model
- DeepSeek-V4-vs-Qwen3.6-35B_comparison.htmlSide-by-side comparison of both models across all benchmark dimensions
README
Qwen3.6-35B vs DSF4 Tool Eval Bench
Evaluation benchmarks comparing Qwen3.6-35B (specifically the Qwen3.6-35B-A3B-UD-Q8_K_XL GGUF quant) against DeepSeek-V4-Flash-DSpark on tool-calling and function-calling tasks.
Test Setup
All benchmarks were run with the following command with THINKING ON using https://github.com/SeraphimSerapis/tool-eval-bench/:
tool-eval-bench --seed 42 --trials 8 --hardmode --base-url http://localhost:8888Contents
| File | Description |
|---|---|
Qwen3.6-35B-A3B-UD-Q8_K_XL.html |
Tool evaluation results for the Qwen3.6-35B model |
DeepSeek-V4-Flash-DSpark.html |
Tool evaluation results for the DeepSeek V4 Flash DSpark model |
DeepSeek-V4-vs-Qwen3.6-35B_comparison.html |
Side-by-side comparison of both models across all benchmark dimensions |
How to View
Open any of the .html files in your browser to see the full benchmark results with interactive tables and visualizations.
License
MIT