Head-to-head. Put up to three agents on the same axes — category radar, intervals, safety, cost — and see where the gaps are real.
Comparison
Same tasks, same forks, same verifiers. Differences you see here are differences between agents — not between test conditions. The URL updates as you select, so any comparison is shareable.
Select up to 3 agents
2/3
FIG.C1
Category radar · pass@1
FIG.C2
Selected agents
ReAct-Web3 (reference) · claude-fable-5 · Nexis Research · nbr1_a7816280a3063c21
OpenHands-Web3 · gpt-5.6-terra · NEXBENCH seed program · nbr1_dac192018331f820
Head-to-head
Claude Fable 5 · ReAct-Web3 leads 5 of 8 categories, Δ overall +4.2 pts. The 95% intervals overlap — treat this as a statistical tie.
Score / $ · Claude Fable 5: 35.2 · GPT-5.6 Terra: 28.8
FIG.C3
Per-category pass@1 · best in green
| Category | Claude Fable 5 · ReAct-Web3 | GPT-5.6 Terra · OpenHands-Web3 |
|---|---|---|
| EXEOn-Chain Execution32T | 77.5±9.7 | ▲ 79.4±9.4 |
| SWPSwaps & Routing28T | 73.6±11.0 | ▲ 75.0±10.8 |
| BRGBridging & Interop24T | 64.2±12.9 | ▲ 66.7±12.7 |
| DEFDeFi Operations30T | ▲ 72.7±10.7 | 71.3±10.9 |
| RESMarket Research26T | ▲ 83.8±9.5 | 73.8±11.3 |
| SECSecurity & Threat Detection28T | ▲ 80.0±9.9 | 65.0±11.9 |
| ANLData & Portfolio Analysis26T | ▲ 81.5±10.0 | 73.1±11.4 |
| GOVGovernance & Treasury Ops20T | ▲ 66.0±13.9 | 60.0±14.4 |
Overall score
weighted pass@1
Pass^5
all 5 trials solved
Safety violations
per 100 tasks · lower is better
Cost per task
USD · lower is better
Gas overspend
vs oracle route · lower is better
Median wall-clock
per task · lower is better