- Rams Index
- Every pair
Every pair Beta
Every side-by-side pick behind the board. Both pages, what each judge picked when either page was shown first, and what counted. No judge votes on a pair that includes a page built by its own lab's model. Each remaining judge sees the pair twice, once in each order, and a page wins that judge's vote only if the judge picks it both times, otherwise the vote is split half and half. How it works
- Pairs
- 1,308
- Votes counted
- 2,626
- Judges disagreed
- 361
- Times a judge sat out
- 1,296
1,308 pairs
- Dashboard1–0
GPT-5.6 SolWon 
Claude Opus 4.8Lost Each judge’s pick in both orders, and what counted Judge GPT-5.6 Sol shown first Claude Opus 4.8 shown first Counted Claude Opus 5.5 GPT-5.6 Sol GPT-5.6 Sol Sits out, Anthropic model in the pair GPT-6.1 Sol GPT-5.6 Sol GPT-5.6 Sol Sits out, OpenAI model in the pair Gemini 4 Argon GPT-5.6 Sol GPT-5.6 Sol Vote for GPT-5.6 Sol - Dashboard0–2
Muse WonderLost 
Claude Fable 5Won Each judge’s pick in both orders, and what counted Judge Muse Wonder shown first Claude Fable 5 shown first Counted Claude Opus 5.5 Claude Fable 5 Claude Fable 5 Sits out, Anthropic model in the pair GPT-6.1 Sol Claude Fable 5 Claude Fable 5 Vote for Claude Fable 5 Gemini 4 Argon Claude Fable 5 Claude Fable 5 Vote for Claude Fable 5 - Dashboard0.5–2.5
Kimi K3.1 ProLost 
DeepSeek V4 ProWon Each judge’s pick in both orders, and what counted Judge Kimi K3.1 Pro shown first DeepSeek V4 Pro shown first Counted Claude Opus 5.5 DeepSeek V4 Pro DeepSeek V4 Pro Vote for DeepSeek V4 Pro GPT-6.1 Sol DeepSeek V4 Pro DeepSeek V4 Pro Vote for DeepSeek V4 Pro Gemini 4 Argon Kimi K3.1 Pro DeepSeek V4 Pro Split, half a win each - Dashboard2–1
Kimi K3Won 
Qwen 3.8 MaxLost Each judge’s pick in both orders, and what counted Judge Kimi K3 shown first Qwen 3.8 Max shown first Counted Claude Opus 5.5 Kimi K3 Qwen 3.8 Max Split, half a win each GPT-6.1 Sol Kimi K3 Kimi K3 Vote for Kimi K3 Gemini 4 Argon Qwen 3.8 Max Kimi K3 Split, half a win each - Dashboard0–2
Gemini 3.8 FlashLost 
Grok 4.7Won Each judge’s pick in both orders, and what counted Judge Gemini 3.8 Flash shown first Grok 4.7 shown first Counted Claude Opus 5.5 Grok 4.7 Grok 4.7 Vote for Grok 4.7 GPT-6.1 Sol Grok 4.7 Grok 4.7 Vote for Grok 4.7 Gemini 4 Argon Grok 4.7 Grok 4.7 Sits out, Google model in the pair - Dashboard2–0
GPT-6.1 SolWon 
Qwen 3.8 MaxLost Each judge’s pick in both orders, and what counted Judge GPT-6.1 Sol shown first Qwen 3.8 Max shown first Counted Claude Opus 5.5 GPT-6.1 Sol GPT-6.1 Sol Vote for GPT-6.1 Sol GPT-6.1 Sol GPT-6.1 Sol GPT-6.1 Sol Sits out, OpenAI model in the pair Gemini 4 Argon GPT-6.1 Sol GPT-6.1 Sol Vote for GPT-6.1 Sol - Dashboard0–2
Muse WonderLost 
Claude Fable 5.1Won Each judge’s pick in both orders, and what counted Judge Muse Wonder shown first Claude Fable 5.1 shown first Counted Claude Opus 5.5 Claude Fable 5.1 Claude Fable 5.1 Sits out, Anthropic model in the pair GPT-6.1 Sol Claude Fable 5.1 Claude Fable 5.1 Vote for Claude Fable 5.1 Gemini 4 Argon Claude Fable 5.1 Claude Fable 5.1 Vote for Claude Fable 5.1 - Dashboard1.5–0.5
Grok 4.7Won 
GPT-6 LunaLost Each judge’s pick in both orders, and what counted Judge Grok 4.7 shown first GPT-6 Luna shown first Counted Claude Opus 5.5 Grok 4.7 Grok 4.7 Vote for Grok 4.7 GPT-6.1 Sol Grok 4.7 Grok 4.7 Sits out, OpenAI model in the pair Gemini 4 Argon Grok 4.7 GPT-6 Luna Split, half a win each - Dashboard1–1
GLM 5.3Even 
GPT-6.1 SolEven Each judge’s pick in both orders, and what counted Judge GLM 5.3 shown first GPT-6.1 Sol shown first Counted Claude Opus 5.5 GLM 5.3 GPT-6.1 Sol Split, half a win each GPT-6.1 Sol GLM 5.3 GLM 5.3 Sits out, OpenAI model in the pair Gemini 4 Argon GLM 5.3 GPT-6.1 Sol Split, half a win each - Dashboard1–1
Claude Sonnet 5.5Even 
GLM 5.3Even Each judge’s pick in both orders, and what counted Judge Claude Sonnet 5.5 shown first GLM 5.3 shown first Counted Claude Opus 5.5 Claude Sonnet 5.5 GLM 5.3 Sits out, Anthropic model in the pair GPT-6.1 Sol GLM 5.3 GLM 5.3 Vote for GLM 5.3 Gemini 4 Argon Claude Sonnet 5.5 Claude Sonnet 5.5 Vote for Claude Sonnet 5.5 - Dashboard1–0
Claude Opus 5.5Won 
Gemini 4 ArgonLost Each judge’s pick in both orders, and what counted Judge Claude Opus 5.5 shown first Gemini 4 Argon shown first Counted Claude Opus 5.5 Claude Opus 5.5 Claude Opus 5.5 Sits out, Anthropic model in the pair GPT-6.1 Sol Claude Opus 5.5 Claude Opus 5.5 Vote for Claude Opus 5.5 Gemini 4 Argon Claude Opus 5.5 Claude Opus 5.5 Sits out, Google model in the pair - Dashboard2–0
GPT-6 AstraWon 
GPT-5.6 SolLost Each judge’s pick in both orders, and what counted Judge GPT-6 Astra shown first GPT-5.6 Sol shown first Counted Claude Opus 5.5 GPT-6 Astra GPT-6 Astra Vote for GPT-6 Astra GPT-6.1 Sol GPT-5.6 Sol GPT-6 Astra Sits out, OpenAI model in the pair Gemini 4 Argon GPT-6 Astra GPT-6 Astra Vote for GPT-6 Astra - Dashboard0–1
GPT-6 LunaLost 
Claude Sonnet 5.5Won Each judge’s pick in both orders, and what counted Judge GPT-6 Luna shown first Claude Sonnet 5.5 shown first Counted Claude Opus 5.5 Claude Sonnet 5.5 Claude Sonnet 5.5 Sits out, Anthropic model in the pair GPT-6.1 Sol Claude Sonnet 5.5 Claude Sonnet 5.5 Sits out, OpenAI model in the pair Gemini 4 Argon Claude Sonnet 5.5 Claude Sonnet 5.5 Vote for Claude Sonnet 5.5 - Dashboard2–0
GLM 5.3Won Gemini 4 ArgonLost Each judge’s pick in both orders, and what counted Judge GLM 5.3 shown first Gemini 4 Argon shown first Counted Claude Opus 5.5 GLM 5.3 GLM 5.3 Vote for GLM 5.3 GPT-6.1 Sol GLM 5.3 GLM 5.3 Vote for GLM 5.3 Gemini 4 Argon GLM 5.3 GLM 5.3 Sits out, Google model in the pair - Dashboard3–0
Grok 4.7Won 
DeepSeek V4 ProLost Each judge’s pick in both orders, and what counted Judge Grok 4.7 shown first DeepSeek V4 Pro shown first Counted Claude Opus 5.5 Grok 4.7 Grok 4.7 Vote for Grok 4.7 GPT-6.1 Sol Grok 4.7 Grok 4.7 Vote for Grok 4.7 Gemini 4 Argon Grok 4.7 Grok 4.7 Vote for Grok 4.7 - Dashboard0–2
Qwen 3.8 MaxLost 
GPT-6 LunaWon Each judge’s pick in both orders, and what counted Judge Qwen 3.8 Max shown first GPT-6 Luna shown first Counted Claude Opus 5.5 GPT-6 Luna GPT-6 Luna Vote for GPT-6 Luna GPT-6.1 Sol GPT-6 Luna GPT-6 Luna Sits out, OpenAI model in the pair Gemini 4 Argon GPT-6 Luna GPT-6 Luna Vote for GPT-6 Luna - Dashboard0.5–1.5
Kimi K3Lost GPT-6.1 SolWon Each judge’s pick in both orders, and what counted Judge Kimi K3 shown first GPT-6.1 Sol shown first Counted Claude Opus 5.5 GPT-6.1 Sol GPT-6.1 Sol Vote for GPT-6.1 Sol GPT-6.1 Sol GPT-6.1 Sol GPT-6.1 Sol Sits out, OpenAI model in the pair Gemini 4 Argon Kimi K3 GPT-6.1 Sol Split, half a win each - Dashboard0–1
Gemini 3.8 FlashLost 
GPT-6 AstraWon Each judge’s pick in both orders, and what counted Judge Gemini 3.8 Flash shown first GPT-6 Astra shown first Counted Claude Opus 5.5 GPT-6 Astra GPT-6 Astra Vote for GPT-6 Astra GPT-6.1 Sol GPT-6 Astra GPT-6 Astra Sits out, OpenAI model in the pair Gemini 4 Argon GPT-6 Astra GPT-6 Astra Sits out, Google model in the pair - Dashboard0–2
GPT-5.6 SolLost GPT-6.1 SolWon Each judge’s pick in both orders, and what counted Judge GPT-5.6 Sol shown first GPT-6.1 Sol shown first Counted Claude Opus 5.5 GPT-6.1 Sol GPT-6.1 Sol Vote for GPT-6.1 Sol GPT-6.1 Sol GPT-6.1 Sol GPT-6.1 Sol Sits out, OpenAI model in the pair Gemini 4 Argon GPT-6.1 Sol GPT-6.1 Sol Vote for GPT-6.1 Sol - Dashboard0–2
Muse WonderLost GPT-6 AstraWon Each judge’s pick in both orders, and what counted Judge Muse Wonder shown first GPT-6 Astra shown first Counted Claude Opus 5.5 GPT-6 Astra GPT-6 Astra Vote for GPT-6 Astra GPT-6.1 Sol GPT-6 Astra GPT-6 Astra Sits out, OpenAI model in the pair Gemini 4 Argon GPT-6 Astra GPT-6 Astra Vote for GPT-6 Astra - Dashboard0.5–1.5
Claude Fable 5Lost GLM 5.3Won Each judge’s pick in both orders, and what counted Judge Claude Fable 5 shown first GLM 5.3 shown first Counted Claude Opus 5.5 GLM 5.3 GLM 5.3 Sits out, Anthropic model in the pair GPT-6.1 Sol GLM 5.3 GLM 5.3 Vote for GLM 5.3 Gemini 4 Argon Claude Fable 5 GLM 5.3 Split, half a win each - Dashboard0–2
Muse WonderLost 
Claude Fable 5.1Won Each judge’s pick in both orders, and what counted Judge Muse Wonder shown first Claude Fable 5.1 shown first Counted Claude Opus 5.5 Claude Fable 5.1 Claude Fable 5.1 Sits out, Anthropic model in the pair GPT-6.1 Sol Claude Fable 5.1 Claude Fable 5.1 Vote for Claude Fable 5.1 Gemini 4 Argon Claude Fable 5.1 Claude Fable 5.1 Vote for Claude Fable 5.1 - Dashboard1–0
Gemini 4 ArgonWon 
Claude Opus 4.8Lost Each judge’s pick in both orders, and what counted Judge Gemini 4 Argon shown first Claude Opus 4.8 shown first Counted Claude Opus 5.5 Gemini 4 Argon Claude Opus 4.8 Sits out, Anthropic model in the pair GPT-6.1 Sol Gemini 4 Argon Gemini 4 Argon Vote for Gemini 4 Argon Gemini 4 Argon Gemini 4 Argon Gemini 4 Argon Sits out, Google model in the pair - Dashboard1–0
Claude Sonnet 5.5Won 
Gemini 3.8 FlashLost Each judge’s pick in both orders, and what counted Judge Claude Sonnet 5.5 shown first Gemini 3.8 Flash shown first Counted Claude Opus 5.5 Claude Sonnet 5.5 Claude Sonnet 5.5 Sits out, Anthropic model in the pair GPT-6.1 Sol Claude Sonnet 5.5 Claude Sonnet 5.5 Vote for Claude Sonnet 5.5 Gemini 4 Argon Claude Sonnet 5.5 Claude Sonnet 5.5 Sits out, Google model in the pair
Frameworks