← Back to the board Model weights Every model starts with an equal vote. As its calls resolve, the panel leans toward the ones whose confidence proves well-calibrated and whose recent form holds up — so a model’s weight drifts above or below the dashed equal-vote line. Nothing here is set by hand; it is read straight from the resolved log.
0.70 0.76 0.82 equal vote start 06-29 06-30 07-02 07-03 07-04 07-06 07-07 07-09 07-10 07-11 anthropic:claude-sonnet-4-6 · start: 0.750 anthropic:claude-sonnet-4-6 · 06-29: 0.760 anthropic:claude-sonnet-4-6 · 06-30: 0.761 anthropic:claude-sonnet-4-6 · 07-02: 0.753 anthropic:claude-sonnet-4-6 · 07-03: 0.753 anthropic:claude-sonnet-4-6 · 07-04: 0.753 anthropic:claude-sonnet-4-6 · 07-06: 0.755 anthropic:claude-sonnet-4-6 · 07-07: 0.753 anthropic:claude-sonnet-4-6 · 07-09: 0.750 anthropic:claude-sonnet-4-6 · 07-10: 0.749 anthropic:claude-sonnet-4-6 · 07-11: 0.748 openai:gpt-4o · start: 0.750 openai:gpt-4o · 06-29: 0.777 openai:gpt-4o · 06-30: 0.773 openai:gpt-4o · 07-02: 0.753 openai:gpt-4o · 07-03: 0.748 openai:gpt-4o · 07-04: 0.748 openai:gpt-4o · 07-06: 0.748 openai:gpt-4o · 07-07: 0.748 openai:gpt-4o · 07-09: 0.748 openai:gpt-4o · 07-10: 0.748 openai:gpt-4o · 07-11: 0.748 openrouter:deepseek/deepseek-chat · start: 0.750 openrouter:deepseek/deepseek-chat · 06-29: 0.746 openrouter:deepseek/deepseek-chat · 06-30: 0.747 openrouter:deepseek/deepseek-chat · 07-02: 0.747 openrouter:deepseek/deepseek-chat · 07-03: 0.745 openrouter:deepseek/deepseek-chat · 07-04: 0.744 openrouter:deepseek/deepseek-chat · 07-06: 0.752 openrouter:deepseek/deepseek-chat · 07-07: 0.750 openrouter:deepseek/deepseek-chat · 07-09: 0.748 openrouter:deepseek/deepseek-chat · 07-10: 0.745 openrouter:deepseek/deepseek-chat · 07-11: 0.743 openrouter:google/gemini-2.5-flash · start: 0.750 openrouter:google/gemini-2.5-flash · 06-29: 0.762 openrouter:google/gemini-2.5-flash · 06-30: 0.759 openrouter:google/gemini-2.5-flash · 07-02: 0.747 openrouter:google/gemini-2.5-flash · 07-03: 0.745 openrouter:google/gemini-2.5-flash · 07-04: 0.745 openrouter:google/gemini-2.5-flash · 07-06: 0.754 openrouter:google/gemini-2.5-flash · 07-07: 0.750 openrouter:google/gemini-2.5-flash · 07-09: 0.747 openrouter:google/gemini-2.5-flash · 07-10: 0.744 openrouter:google/gemini-2.5-flash · 07-11: 0.743 openrouter:x-ai/grok-4.3 · start: 0.750 openrouter:x-ai/grok-4.3 · 06-29: 0.768 openrouter:x-ai/grok-4.3 · 06-30: 0.768 openrouter:x-ai/grok-4.3 · 07-02: 0.752 openrouter:x-ai/grok-4.3 · 07-03: 0.748 openrouter:x-ai/grok-4.3 · 07-04: 0.748 openrouter:x-ai/grok-4.3 · 07-06: 0.755 openrouter:x-ai/grok-4.3 · 07-07: 0.754 openrouter:x-ai/grok-4.3 · 07-09: 0.750 openrouter:x-ai/grok-4.3 · 07-10: 0.747 openrouter:x-ai/grok-4.3 · 07-11: 0.746 Current standings Model Weight vs equal Resolved calls anthropic:claude-sonnet-4-60.75 ±0.00 712 openai:gpt-4o0.75 ±0.00 232 openrouter:x-ai/grok-4.30.75 ±0.00 808 openrouter:deepseek/deepseek-chat0.74 −0.01 808 openrouter:google/gemini-2.5-flash0.74 −0.01 808
Weight is a model’s vote share before its own confidence is applied. A model only moves off equal once it has a genuine, sustained record — a couple of lucky calls barely shift it.