Hermes tool-contract hard v1: 50 cases. Primary score is tool/state pass rate. tool_state=70%, strict=70%, final_text=96%, schema_valid=100%, required_tool=78%, hallucinated_tool=0%, forbidden_tool=0%, recovery=100.0%. avg_tool_calls=2.98, avg_turns=3.94, output_tps=35.0, input_tokens=137305, output_tokens=33051. non_text_failures: final_state_wrong=11, max_tool_calls=2, missing_required_tool=11, too_many_tool_calls=2. strict_failure_breakdown: final_answer_missing_expected_text=2, final_state_wrong=11, max_tool_calls=2, missing_required_tool=11, too_many_tool_calls=2. Local OMLX AgentWorld oQ4 run.
Artifact/log: -OMLX local
Qwen AgentWorld 35B A3B oQ4 MLX
Local OpenAI-compatible OMLX provider; oQ4 MLX quantization.
Score Summary
Overall eligibility requires eligible data in all four categories.
Overall
Incomplete for overall leaderboards
Agentic Tool Use
2/2 benchmarks covered
Agentic Coding
0/0 benchmarks covered
Long-Term Tasks
0/0 benchmarks covered
Speed
0/0 benchmarks covered
Latest Results
2 eval runs
Hermes tool-contract v0: 100 cases. Primary score is tool/state pass rate. tool_state=85%, strict=81%, final_text=91%, schema_valid=100%, required_tool=90%, hallucinated_tool=0%, forbidden_tool=0%, recovery=66.7%. avg_tool_calls=1.31, avg_turns=2.25, output_tps=39.3, input_tokens=108131, output_tokens=39972. non_text_failures: final_state_wrong=4, max_tool_calls=5, missing_required_tool=10. strict_failure_breakdown: final_answer_missing_expected_text=9, final_state_wrong=4, max_tool_calls=5, missing_required_tool=10. Local OMLX AgentWorld oQ4 run.
Artifact/log: -Historical Runs
2 eval runs
Hermes tool-contract hard v1: 50 cases. Primary score is tool/state pass rate. tool_state=70%, strict=70%, final_text=96%, schema_valid=100%, required_tool=78%, hallucinated_tool=0%, forbidden_tool=0%, recovery=100.0%. avg_tool_calls=2.98, avg_turns=3.94, output_tps=35.0, input_tokens=137305, output_tokens=33051. non_text_failures: final_state_wrong=11, max_tool_calls=2, missing_required_tool=11, too_many_tool_calls=2. strict_failure_breakdown: final_answer_missing_expected_text=2, final_state_wrong=11, max_tool_calls=2, missing_required_tool=11, too_many_tool_calls=2. Local OMLX AgentWorld oQ4 run.
Artifact/log: -Hermes tool-contract v0: 100 cases. Primary score is tool/state pass rate. tool_state=85%, strict=81%, final_text=91%, schema_valid=100%, required_tool=90%, hallucinated_tool=0%, forbidden_tool=0%, recovery=66.7%. avg_tool_calls=1.31, avg_turns=2.25, output_tps=39.3, input_tokens=108131, output_tokens=39972. non_text_failures: final_state_wrong=4, max_tool_calls=5, missing_required_tool=10. strict_failure_breakdown: final_answer_missing_expected_text=9, final_state_wrong=4, max_tool_calls=5, missing_required_tool=10. Local OMLX AgentWorld oQ4 run.
Artifact/log: -