1551 GPT-4o mini OpenRouter harness, openai/gpt-4o-mini, Azure, public 90-day aggregate · OpenRouter benchmark exports 43.1% 1.5k 1552 Qwen2.5 VL 72B Instruct OpenRouter harness, qwen/qwen2.5-vl-72b-instruct, auto-routing, public 90-day aggregate · OpenRouter benchmark exports 42.9% 1.7k 1553 GPT-4o mini OpenRouter harness, openai/gpt-4o-mini, OpenAI, public 90-day aggregate · OpenRouter benchmark exports 42.8% 1.5k 1554 GPT 4 1106 Preview · Epoch AI Benchmarking Hub data 42.4%
— Not measured 1555 GPT-4-0125-preview · Epoch AI Benchmarking Hub data 42.3%
— Not measured 1556 Mistral Small 3.1 · Epoch AI Benchmarking Hub data 41.9%
— Not measured 1557 Qwen Turbo · Epoch AI Benchmarking Hub data 41.8%
— Not measured 1558 Llama 3.2 90B Vision Instruct · Epoch AI Benchmarking Hub data 41%
— Not measured 1559 Qwen2 72B Instruct · Epoch AI Benchmarking Hub data 40.8%
— Not measured 1560 Claude 3 Sonnet · Epoch AI Benchmarking Hub data 40.6%
— Not measured 1561 Gemma 3 12B IT OpenRouter harness, google/gemma-3-12b-it, DeepInfra, public 90-day aggregate · OpenRouter benchmark exports 40.6% 1.9k 1562 Llama 3 70B Instruct · Epoch AI Benchmarking Hub data 40.6%
— Not measured 1563 gemini-1.5-flash-001 · Epoch AI Benchmarking Hub data 40.4%
— Not measured 1564 Gemma 3 12B IT · Epoch AI Benchmarking Hub data 39.5%
— Not measured 1565 Mistral Large 2402 · Epoch AI Benchmarking Hub data 38.8%
— Not measured 1566 Gemma 4 26B A4B OpenRouter harness, google/gemma-4-26b-a4b-it-20260403, Venice, public 90-day aggregate · OpenRouter benchmark exports 38.6% 21k 1567 Nova Micro OpenRouter harness, amazon/nova-micro-v1, Amazon Bedrock, public 90-day aggregate · OpenRouter benchmark exports 38.6% 932 1568 Claude Haiku 3.5 · Epoch AI Benchmarking Hub data 38.1%
— Not measured 1569 Gemma 3 12B IT OpenRouter harness, google/gemma-3-12b-it, NextBit, public 90-day aggregate · OpenRouter benchmark exports 38% 2.3k 1570 Qwen3 1.7B · Epoch AI Benchmarking Hub data 38%
— Not measured 1571 Gemma 4 31B IT OpenRouter harness, google/gemma-4-31b-it-20260402, Io Net, public 90-day aggregate · OpenRouter benchmark exports 37.9% 5.1k 1572 GPT-4o mini · Epoch AI Benchmarking Hub data 37.7%
— Not measured 1573 Gemma 2 27B IT · Epoch AI Benchmarking Hub data 36.5%
— Not measured 1574 Claude 3 Haiku · Epoch AI Benchmarking Hub data 36.3%
— Not measured 1575 GPT 4 · Epoch AI Benchmarking Hub data 35.7%
— Not measured 1576 Qwen2.5 7B Instruct · Epoch AI Benchmarking Hub data 35.5%
— Not measured 1577 DeepSeek V4 Flash OpenRouter harness, deepseek/deepseek-v4-flash-20260423, Reka, public 90-day aggregate · OpenRouter benchmark exports 34.7% 33k 1578 Claude 2 · Epoch AI Benchmarking Hub data 34.7%
— Not measured 1579 DeepSeek R1 Distill Qwen 1.5B · Epoch AI Benchmarking Hub data 33.6%
— Not measured 1580 Claude 2.1 · Epoch AI Benchmarking Hub data 33%
— Not measured 1581 gemini-1.5-flash-8b-001 · Epoch AI Benchmarking Hub data 33%
— Not measured 1582 DBRX-instruct · Epoch AI Benchmarking Hub data 32.9%
— Not measured 1583 Yi 1.5 34B Chat · Epoch AI Benchmarking Hub data 32%
— Not measured 1584 Qwen1.5-32B-Chat · Epoch AI Benchmarking Hub data 30.7%
— Not measured 1585 Qwen3 1.7B none · Epoch AI Benchmarking Hub data 30.6%
— Not measured 1586 Mixtral 8x7B Instruct v0.1 · Epoch AI Benchmarking Hub data 30.6%
— Not measured 1587 Llama 3.1 8B Instruct · OpenEvals Leaderboard data · self-reported Reported by the model's developer. The source relaying it did not verify it. Values from 2 sources 30.4%
— Not measured 1588 Open Mistral Nemo · Epoch AI Benchmarking Hub data 29.9%
— Not measured 1589 open-mixtral-8x7b · Epoch AI Benchmarking Hub data 29.8%
— Not measured 1590 Llama 3.1 8B Instruct OpenRouter harness, meta-llama/llama-3.1-8b-instruct, Groq, public 90-day aggregate · OpenRouter benchmark exports 29.2% 37k 1591 Qwen1.5-72B-Chat · Epoch AI Benchmarking Hub data 28.8%
— Not measured 1592 Llama 3.2 3B OpenRouter harness, meta-llama/llama-3.2-3b-instruct, Cloudflare, public 90-day aggregate · OpenRouter benchmark exports 28.6% 1.6k 1593 GPT-3.5 Turbo · Epoch AI Benchmarking Hub data 28%
— Not measured 1594 Phi-3 Medium 128K Instruct · Epoch AI Benchmarking Hub data 27.6%
— Not measured 1595 Gemma 2 9B IT · Epoch AI Benchmarking Hub data 27.5%
— Not measured 1596 Ministral 8B · Epoch AI Benchmarking Hub data 27.1%
— Not measured 1597 Gemma 3 4B IT OpenRouter harness, google/gemma-3-4b-it, Parasail, public 90-day aggregate · OpenRouter benchmark exports 27.1% 1.8k 1598 Llama 3.1 8B Instruct OpenRouter harness, meta-llama/llama-3.1-8b-instruct, auto-routing, public 90-day aggregate · OpenRouter benchmark exports 27.1% 13k 1599 Llama 3.1 8B Instruct OpenRouter harness, meta-llama/llama-3.1-8b-instruct, CoreWeave, public 90-day aggregate · OpenRouter benchmark exports 26.7% 53k 1600 Llama-2-70b-chat-hf · Epoch AI Benchmarking Hub data 26.3%
— Not measured