Gemini 4 Argon 空降 Debate Benchmark 第 4 名

Lech Mazur @LechMazur原推发布于 2026.10.15

在 X 上打开原推

Gemini 4 Argon and Meta Muse Wonder enter the Debate Benchmark top 5 among current models. Gemini 4 Argon (high) debuts at #4 with 1676 points, gaining 190 over Gemini 3.1 Pro Preview. Meta Muse Wonder 1.0 (high) enters at #5 with 1655 points, up 62 over Muse Spark 1.3. Mistral Large 4 (high) delivers the biggest predecessor gain in this update: 1177 → 1462 (+285) over Mistral Large 3. It also beats Mistral Medium 3.5 by 126 points, though it remains well behind the leaders. The benchmark now includes 7,922 rated debates across 683 topics, backed by 24,324 judge decisions. Debate Benchmark tests how well models defend a position through sustained, adversarial, multi-turn opposition. It demands broad knowledge, factual accuracy under pressure, sharp rebuttals, and arguments that hold together round after round. Every matchup runs twice on the same motion, with PRO and CON swapped to control for side advantage. Three judges from distinct model families independently evaluate each debate’s winner and margin. More info: https://github.com/lechmazur/debate