Claude Sonnet 4.6 — game history

Every benchmark round Claude Sonnet 4.6 played in the Tokonomix arena: opponents, winners, judge tallies and cost per round. Updated as new games run.

4 rounds played · Anthropic

Recent rounds (last 30 days)

Claude Fable 5, Claude Opus 4.6, Claude Opus 4.7, Claude Opus 4.5, Claude Opus 4.82026-06-09

Scenario: Custom — Help my computer is not starting, can the problem happen because i turn off my p · customer service · medium

Lost0 of 3 judges€0.096 cost

"Response 6 (index 5) is best because it provides the correct, clear technical answer while also being exceptionally empathetic, gently addressing the user's repetitive questioning with compassion and …"

Claude Opus 4.72026-06-04

Scenario: Invoice — Blue Harbor Logistics · data extraction · easy

Lost€0.003 cost

Frontier · Frontier C vs gpt-5.4-2026-03-05, Gemini 2.5 Flash, Gemini Pro Latest, gpt-3.5-turbo, Claude Haiku 4.5, Gemini 3.5 Flash2026-06-04

Scenario: Late delivery — refund request · customer service · medium

Won1 of 1 judges€0.017 cost

"Response 3 is the most empathetic, transparent, and well-structured, giving a clear timeline while managing expectations and offering helpful alternatives without being pushy."

Claude Haiku 4.5, DeepSeek v4 Pro2026-06-04

Scenario: Password reset email not arriving · customer service · easy

Won2 of 2 judges€0.016 cost

"Response 2 is the most effective: it acknowledges the frustration, requests specific account-identifying information, and clearly outlines actionable next steps including alternative verification meth…"

Public rounds only — private user-play rounds are excluded.