Which model leads AgentWorldBench - Terminal?

Q: Which model leads AgentWorldBench - Terminal?

Claude Opus 4.8 tops the board at 59.18

AgentWorldBench - Terminal Leaderboard

Pro Mode — raw scores, one benchmark under the lensthird-party eval harness

The field · 14 measured

41.6259.18

14 models measured, most on the third-party eval harness. Claude Opus 4.8 tops the board at 59.18.

14 measured·Lifecycle Updated Oct 8, 2026Score basis

#◆Vendor$/M in$/M out

Leader — ignites whiteField — fill cools with scoreIndependently verified only0 of 14Bars show 41.62–59.18 · ◆ solid = first-party or better · outlined = arm's-length

01Claude Opus 4.8Anthropic59.18$5.0/M$25.0/M$0.253

02Qwen AgentWorld 397B A17BAlibaba57.73———

03Claude Opus 4.6Anthropic57.51$5.0/M$25.0/M$0.261

04Claude Sonnet 4.6Anthropic56.98$3.0/M$15.0/M$0.158

05Qwen3.5 397B A17BAlibaba55.3$0.60/M$3.6/M$0.038

06Qwen AgentWorld 35B A3BAlibaba53.96———

07GPT-5.4OpenAI53.69$2.5/M$15.0/M$0.163

08Kimi K2.6Moonshot52.54$0.95/M$4.0/M$0.047

09Gemini 3.1 ProGoogle52.47$2.0/M$12.0/M$0.133

10DeepSeek-V4-ProDeepSeek51.26$0.43/M$0.87/M$0.013

11Qwen3.6 PlusAlibaba50.58$0.50/M$3.0/M$0.035

12GLM-5.1Z.ai47.32$1.4/M$4.4/M$0.061

13Qwen3.5 35B A3BAlibaba46.13$0.25/M$2.0/M$0.024

14MiniMax M2.7MiniMax41.62$0.21/M$0.84/M$0.013

◆ solid = independently verified · aggregator attested · vendor attributed — outlined = cross-reference · source-attributed. Dimmed ◆ = confidence inferred. “+N alts” = alternate disclosures in the drawer with why-demoted. Prices per million tokens; $/point = input price ÷ score. Every row opens its model dossier.

Verification: 14 models scored · 0 independently verified · 9 vendor cross-reference · 5 vendor-reported · 0 with source disagreement. How these tiers are assigned