List your company

Multi-turn function calling on BFCL BASE and LONG CONTEXT multi-turn v3

41.88Avg@4 Success RateHINT-SD-Multi
25.302429.606233.9138.2138May 18, 2026

Evaluation results

MethodLinks
HINT-SD-Multibackbone=Qwen3-4B-Instruct-2507, distill_strategy=multiple selected failure-relevant steps, rollouts=42026.0541.8848.75
HINT-SD-Singlebackbone=Qwen3-4B-Instruct-2507, distill_strategy=first failure-relevant step, rollouts=42026.0536.2543.13
GRPObackbone=Qwen3-4B-Instruct-2507, rollouts=42026.0531.5641.25
SDPObackbone=Qwen3-4B-Instruct-2507, rollouts=42026.0530.7840
SFTmode=supervised fine-tuning, backbone=Qwen3-4B-Instruct-25072026.0528.4438.13
OpenClaw-RLbackbone=Qwen3-4B-Instruct-2507, rollouts=42026.0528.2845
Initialmode=zero-shot, backbone=Qwen3-4B-Instruct-25072026.0525.9436.25