List your company

Tool Use on Unified evaluation benchmark D Single-Hop (test)

96.1SP ScoreQwen3-8B (Upper Bound)
44.93258.21671.584.784Apr 13, 2026

Evaluation results

MethodLinks
Qwen3-8B (Upper Bound)Setting=GT Setting2026.0496.196.128.752.3
UniToolCallSetting=Hybrid-20 Setting2026.0492.992.927.148.6
Qwen3-32BSetting=Hybrid-20 Setting2026.0472.672.623.743.8
Gemini 3 Flash PreviewSetting=Hybrid-20 Setting2026.0468.368.325.246.3
Qwen3-8B (Vanilla)Setting=Hybrid-20 Setting2026.0466.966.919.819.8
Claude 4.6 SonnetSetting=Hybrid-20 Setting2026.0458.858.824.943.1
Kimi-K2-InstructSetting=Hybrid-20 Setting2026.0455.955.921.337.1
GPT-5.2 InstantSetting=Hybrid-20 Setting2026.0450.950.923.239.2
DeepSeek-V3.2Setting=Hybrid-20 Setting2026.0446.946.918.832.8