Code Generation on LiveCode Bench (Performance % and Total Tokens)
1.37Total TokensAgentFmw
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| AgentFmwBackbone=GPT-OSS:120B2026.04 | 1.37 | — | |
| TopoDIMBackbone=GPT-OSS:120B2026.04 | 1.39 | — | |
| G-DesignerBackbone=GPT-OSS:120B2026.04 | 2.63 | — | |
| LongGraphBackbone=GPT-OSS:120B2026.04 | 2.93 | — | |
| GTDBackbone=GPT-OSS:120B2026.04 | 3.12 | — | |
| AutoGenBackbone=GPT-OSS:120B2026.04 | 3.42 | — | |
| GPTSwarmBackbone=GPT-OSS:120B2026.04 | 4.9 | — | |
| LLM-DebateBackbone=GPT-OSS:120B2026.04 | 8.81 | — | |
| Agent Q-MixBackbone=GPT-OSS:120B2026.04 | 312 | — | |
| LobsterBackbone=GPT-OSS:120B2026.04 | 827 | — |