List your company

Function Calling on BFCL V4 (Custom Metrics Set)

62.3Multi-Turn Success RateSeed 2.0
1.77217.48633.248.914Apr 20, 2026Apr 27, 2026May 5, 2026May 13, 2026May 20, 2026May 28, 2026Jun 5, 2026

Evaluation results

MethodLinks
Seed 2.0Model Category=Frontier Proprietary Models2026.0462.39257.88982.276.67573.4
Claude Sonnet-4.5Model Category=Frontier Proprietary Models2026.0461.4816588.781.168.886.673.2
Gemini-3 ProModel Category=Frontier Proprietary Models2026.0460.88061.790.783.168.885.672.5
Agent-World-14BModel Category=Open-Source Environment Scaling Methods, Parameter Scale=14B2026.0453.95323.982.379.393.88155.8
4B-ThinkingCategory=Open-source Baselines2026.0652612.3738.0271.06—79.2338.61
235B-A22B-InstructCategory=Open-source Baselines2026.0650.1244.527.190.183.2—82.3853.54
Turn-level TRUSTBase Model=Qwen3-4B-Thinking, Training Level=Turn-level2026.0649.62324.9589.2384.09—84.8448.04
GPT-5.2 HighModel Category=Frontier Proprietary Models2026.0448.575.545.881.970.47588.762.9
EnvScaler-8BModel Category=Open-Source Environment Scaling Methods, Parameter Scale=8B2026.0447.12321.988.582.293.874.647.6
Qwen3-235B-A22BModel Category=Open-Source Foundation Models, Parameter Scale=235B2026.0445.45423.937.468.987.581.747.9
MAESTRO2026.0544.62——87.3382.38——78.09
Agent-World-8BModel Category=Open-Source Environment Scaling Methods, Parameter Scale=8B2026.0444.54721.783.379.693.880.251.4
GPT-5.22026.0543.75——78.2967.14——68.58
Qwen3-32BModel Category=Open-Source Foundation Models, Parameter Scale=32B2026.0443.32615.790.38281.382.446.7
SAGEBase Model=Qwen3-4B-Thinking2026.0641.53.513.0286.7180.49—79.2341.16
8B-ThinkingCategory=Open-source Baselines2026.0637.751214.6287.5880.53—79.0741.58
AWM-14BModel Category=Open-Source Environment Scaling Methods, Parameter Scale=14B2026.0437.61019.890.281.57579.442.4
DeepSeek-V3.2-685BModel Category=Open-Source Foundation Models, Parameter Scale=685B2026.0437.469.554.234.953.737.593.254.1
Qwen3-14BModel Category=Open-Source Foundation Models, Parameter Scale=14B2026.0436.9419.89082.481.379.441
SAGEBase Model=Qwen3-8B-Thinking2026.0636.379.521.7288.1580.83—79.642.99
Gemini-2.5-Flash2026.0536.25——84.9674.39——72.88
Qwen3-8BModel Category=Open-Source Foundation Models, Parameter Scale=8B2026.0435.4717.690.280.981.377.240.4
AWM-8BModel Category=Open-Source Environment Scaling Methods, Parameter Scale=8B2026.0434.99.515.790.280.593.873.940
30B-A3B-InstructCategory=Open-source Baselines2026.063022.517.6385.7777.94—79.941
Traj.-level TRUSTBase Model=Qwen3-8B-Base, Training Level=Traj.-level2026.0629.882730.5782.2775.57—79.3344.48
AUQBase Model=Qwen3-4B-Thinking2026.0623.5320.4377.0862.4—81.3735.21
Turn-level GRPOBase Model=Qwen3-4B-Thinking, Training Level=Turn-level2026.0622.75324.5280.6186.33—82.3838.98
Traj.-level CM2Base Model=Qwen3-8B-Base, Training Level=Traj.-level2026.0622.52221.9480.1274.58—71.4638.15
TOUCAN-7BModel Category=Open-Source Environment Scaling Methods, Parameter Scale=7B2026.0417.82118.58173.981.378.636.6
Claude-Opus-4.52026.0516.12——89.6576.02——72.14
AUQBase Model=Qwen3-8B-Thinking2026.06101415.784.4675.42—40.1429.08
Simulator-8BModel Category=Open-Source Environment Scaling Methods, Parameter Scale=8B2026.044.117.5647.644.631.387.323.9