--- title: "LLM Leaderboard 2026: Compare GPT, Claude, Gemini & More" description: "Compare leading AI models across benchmarks and see which LLMs are best for support, reasoning, coding, and business automation." canonical: "https://www.whisperchat.ai/tools/llm-leaderboard-compare-model" last-updated: "2026-09-27" source: "https://www.whisperchat.ai/tools/llm-leaderboard-compare-model" --- # LLM Leaderboard 2026: Compare GPT, Claude, Gemini & More > Compare leading AI models across benchmarks and see which LLMs are best for support, reasoning, coding, and business automation. Compare the latest Large Language Models across multiple benchmarks and performance metrics Total Models 71 Latest models included Benchmarks 8 Comprehensive tests Top Score 99.0% Kimi K2.5 (HumanEval) Fastest Model 250 TPS (GPT-4o mini Realtime) Filters & Search Model Rankings - 71 Models # Organization Category MMLU GPQA MMMU HellaSwag HumanEval BBHard GSM8K MATH Cost/1K TPS Context Trend 1 Gemini 3.1 Pro 2026-02 Top GPQABest ARC-AGI-2Best Value Frontier New Google multimodal 92.2% 93.1% 94.3% 85.2% N/A N/A N/A N/A 96.1% $0.002 60 2M 2 GPT-5.4 2026-03 Best Computer UseTop OSWorld New OpenAI reasoning 91.6% 93% 92.8% 83.5% N/A N/A N/A N/A 97% $0.0025 70 1M 3 Claude Opus 4.6 2026-02 Best SWE-BenchTop Coding128K Output New Anthropic coding 91.3% 92.4% 91.3% 85.1% N/A N/A N/A N/A 95.2% $0.005 45 1M 4 Grok 4 2026-01 Best HLEMulti-AgentReal-time X Data New xAI reasoning 90.2% 92.7% 84.6% N/A N/A N/A N/A N/A 93.3% $0.002 75 128K 5 Kimi K2.5 2026-01 Open SourceBest HumanEval OpenTop SWE-Bench Open New Moonshot AI coding 89.5% 92% 87.6% N/A N/A 99% N/A N/A 98% $0.0015 85 262K 6 Claude Sonnet 4.6 2026-02 Best GDPval-AANear-Opus Performance New Anthropic reasoning 88.8% 91% 88.5% 82% N/A N/A N/A N/A 93.5% $0.003 80 1M 7 o1 2024-09 Top MMLUPremium OpenAI reasoning 88.4% 92.3% 78% N/A N/A N/A N/A N/A 94.8% $0.06 15 128K 8 GLM-5 2026-01 Open SourceMIT LicenseBest Chatbot Arena Open New Zhipu AI coding 88% 92% 86% N/A N/A 96.5% N/A N/A 94.5% $0.001 80 200K 9 Qwen 3.5 397B 2026-02 Open SourceApache 2.0Top Open GPQA New Alibaba reasoning 87.7% 91.8% 88.4% N/A N/A 94.2% N/A N/A 96.5% $0.001 90 256K 10 MiniMax M2.5 2026-01 Open SourceBest SWE-Bench Open New MiniMax coding 87.5% 90.8% 85.2% N/A N/A 94.5% N/A N/A 96% $0.0012 75 128K 11 DeepSeek R1 2025-01 Best MATHBest Reasoning New DeepSeek reasoning 86.5% 90.8% 71.5% N/A N/A N/A N/A N/A 97.3% $0.008 65 64K 12 o3-mini 2024-12 Fastest TPSBest HumanEval New OpenAI coding 86% 86% 75% N/A N/A 97% N/A N/A N/A $0.02 189 128K 13 Claude 3.7 Sonnet 2024-11 Best MATH New Anthropic reasoning 85.5% 86.1% 84.8% 75% N/A N/A N/A N/A 96.2% $0.02 65 200K 14 o4-mini 2024-12 Best MATH New OpenAI reasoning 85.2% N/A 81.4% 81.6% N/A N/A N/A N/A 92.7% $0.02 85 128K 15 Gemini 2.5 Pro 2024-12 LatestHigh GPQA New Google multimodal 85.2% 89.8% 84% 81.7% N/A N/A N/A N/A N/A $0.02 55 1M 16 o3 2024-12 Top MATH New OpenAI reasoning 85% N/A 83.3% 82.9% N/A N/A N/A N/A 88.9% $0.06 18 128K 17 o1-preview 2024-09 Preview OpenAI reasoning 84.9% 90.8% 78.3% N/A N/A N/A N/A N/A 85.5% $0.045 20 128K 18 DeepSeek V3.2 2025-12 Open SourceMIT LicenseBest Budget New DeepSeek coding 84.2% 90.5% 72.1% N/A N/A 91.5% N/A N/A 95% $0.00014 95 64K 19 DeepSeek V3 (0324) 2025-03 Open SourceUpdatedMIT License New DeepSeek coding 84.2% 89% 68.4% N/A N/A 87.5% N/A N/A 92% $0.004 95 64K 20 DeepSeek R1 Zero 2025-01 Open SourceRL-Only TrainingMIT License New DeepSeek reasoning 83.8% 88.4% 67% N/A N/A N/A N/A N/A 95.9% $0.005 60 64K 21 Llama 4 Behemoth 2025-07 Open SourceLargest LlamaTop MMLU Open New Meta reasoning 83.4% 91.5% 74.2% 78.3% N/A N/A N/A N/A 89.5% $0.01 25 256K 22 Claude 3.5 Sonnet 2024-06 User's ChoiceBest GSM8K Anthropic reasoning 82.3% 88.7% 59.4% 68.3% 89% 92% 93.1% 96.4% 71.1% $0.015 170 200K 23 GPT-4o 2024-05 Least LatencyMultimodal OpenAI multimodal 82.2% 88.7% 53.6% 69.1% 94.2% 90.2% 91.3% 89.8% 76.6% $0.015 85 128K 24 o1-mini 2024-09 Good Coding OpenAI coding 81.9% 85.2% 60% N/A N/A 92.4% N/A N/A 90% $0.025 45 128K 25 Gemini 2.0 Flash 2024-12 FastGood Performance New Google multimodal 81.8% 87% 59% N/A N/A 91% N/A N/A 90% $0.01 110 1M 26 Claude Opus 4 2024-12 LatestPremium New Anthropic reasoning 81% 88.8% 83.3% 76.5% N/A N/A N/A N/A 75.5% $0.045 45 200K 27 Gemini 2.0 Flash Thinking 2025-02 Extended ThinkingBest Budget Reasoning New Google reasoning 80.7% 85% 70.3% 73.8% N/A N/A N/A N/A 93.5% $0.0035 50 1M 28 Grok 3 Mini 2025-02 Extended ThinkingCost-Effective New xAI reasoning 80.7% 83% 69.7% N/A N/A N/A N/A N/A 89.5% $0.003 100 128K 29 DeepSeek V3 2024-12 Open SourceGood MATH New DeepSeek coding 80.1% 88.5% 59.1% N/A N/A 82.6% N/A N/A 90.2% $0.004 95 64K 30 Claude 3.5 Sonnet v2 2025-02 Computer UseTop SWE-BenchUpgraded New Anthropic reasoning 79.3% 88.7% 65% 70.7% N/A 93.7% N/A N/A 78.3% $0.015 75 200K 31 Llama 3.1 405B 2024-07 Open SourceLargest Open Meta reasoning 78.9% 88.6% 51.1% 64.5% 87% 89% 81.3% 96.8% 73.8% $0.015 35 128K 32 Claude Sonnet 4 2024-12 LatestHigh GPQA New Anthropic reasoning 78.8% 86.5% 83.8% 74.4% N/A N/A N/A N/A 70.5% $0.025 60 200K 33 Qwen 2.5-Max 2025-02 MoE ArchitectureTop Chinese Open New Alibaba reasoning 78.1% 87% 52.5% N/A N/A 88% N/A N/A 85% $0.0016 95 128K 34 GPT-4 Turbo 2024-04 Highly PreferredBalanced OpenAI reasoning 77.6% 86.5% 48% 63.1% 94.2% 90.2% 87.6% 91% 72.2% $0.03 45 128K 35 Llama 3.1 Nemotron 70B 2025-01 Open SourceRLHF TunedTop Arena New NVIDIA reasoning 77.5% 85% 55.8% N/A N/A 90% N/A N/A 79% $0.0035 70 128K 36 GPT-4o (2025) 2025-05 UpdatedBest VoiceImage Gen New OpenAI multimodal 77.2% 89.5% 55% 70.2% N/A 91.5% N/A N/A 80% $0.0125 90 128K 37 Claude 3 Opus 2024-03 PremiumComplete Benchmarks Anthropic reasoning 77.2% 86.8% 50.4% 59.4% 95.4% 84.9% 86.8% 95% 60.1% $0.045 45 200K 38 GPT-4.1 2024-11 Latest GPT New OpenAI reasoning 77.1% 90.2% 66.3% 74.8% N/A N/A N/A N/A N/A $0.04 50 128K 39 Gemini 2.0 Pro Experimental 2024-12 ExperimentalGood MATH New Google multimodal 77.1% 79.1% 64.7% 72.7% N/A N/A N/A N/A 91.8% $0.015 60 1M 40 Qwen 2.5 72B 2025-01 Open SourceApache 2.0Best Open Coding New Alibaba coding 76.4% 86.1% 49% N/A N/A 87.2% N/A N/A 83.1% $0.0009 88 128K 41 Claude 3.7 Sonnet (Normal) 2024-11 Balanced New Anthropic reasoning 76.3% 83.2% 68% 71.8% N/A N/A N/A N/A 82.2% $0.015 85 200K 42 Phi-4 2025-01 Open SourceBest-in-Class 14BSTEM Strong New Microsoft reasoning 75.9% 84.8% 56.1% N/A N/A 82.6% N/A N/A 80.4% $0.0007 120 16K 43 Llama 4 Maverick 2024-12 Open Source New Meta reasoning 75.9% 84.6% 69.8% 73.4% N/A N/A N/A N/A N/A $0.005 85 128K 44 Llama 3.3 70B 2024-10 Open SourceGood Coding Meta coding 75.5% 86% 50.5% N/A N/A 88.4% N/A N/A 77% $0.006 90 128K 45 Mistral Large 2 2025-01 Open WeightsMultilingualFunction Calling New Mistral AI reasoning 75.4% 84% 49.6% N/A N/A 92% N/A N/A 76% $0.006 80 128K 46 Mistral Medium 3 2025-05 EnterpriseMultilingualNew New Mistral AI reasoning 75.2% 83.5% 51% N/A N/A 90% N/A N/A 76.5% $0.004 95 128K 47 GPT-4.1 mini 2024-11 Cost-Effective New OpenAI reasoning 75.1% 87.5% 65% 72.7% N/A N/A N/A N/A N/A $0.015 95 128K 48 Grok-2 2024-08 Good Coding xAI coding 74.8% 87.5% 56% 66.1% N/A 88.4% N/A N/A 76.1% $0.01 75 128K 49 Grok 3 2024-12 Latest New xAI reasoning 74.3% N/A 75.4% 73.2% N/A N/A N/A N/A N/A $0.012 70 128K 50 Gemini 1.5 Pro 2024-02 Largest ContextComplete Benchmarks Google multimodal 73.6% 81.9% 46.2% 62.2% 92.5% 71.9% 84% 91.7% 58.5% $0.0125 38 2M 51 Gemini 2.5 Flash Lite 2024-12 Latest New Google multimodal 71.8% 84.5% 66.7% 72.9% N/A N/A N/A N/A 63.1% $0.01 75 1M 52 GPT-4 2023-03 Most ExpensiveClassic OpenAI reasoning 71.4% 86.4% 35.7% 56.8% 95.3% 67% 83.1% 92% 52.9% $0.18 25 8K 53 Claude 3.5 Haiku (2025) 2025-04 UpdatedFastest ClaudeComputer Use New Anthropic conversation 70.3% 73.5% 43.2% N/A N/A 90.5% N/A N/A 74% $0.004 150 200K 54 Llama 3.2 90B 2024-09 Open Source Meta reasoning 69.6% 86% 46.7% 60.3% N/A N/A N/A 86.9% 68% $0.008 80 128K 55 Command R+ (2025) 2025-03 RAG OptimizedTool UseEnterprise New Cohere reasoning 69.4% 82.3% 46% N/A N/A 80.5% N/A N/A 68.9% $0.0025 85 128K 56 Claude 3 Sonnet 2024-03 BalancedComplete Benchmarks Anthropic reasoning 69.1% 79% 40.4% 53.1% 89% 73% 82.9% 92.3% 43.1% $0.012 90 200K 57 Gemini 1.5 Flash 2024-05 FastComplete Benchmarks Google multimodal 68.6% 78.9% 39.5% 56.1% 81.3% 67.5% 89.2% 68.8% 67.7% $0.008 95 1M 58 Mistral Small 3 2025-03 Open WeightsUltra EfficientApache 2.0 New Mistral AI conversation 68.1% 81.5% 42% N/A N/A 83% N/A N/A 66% $0.001 130 32K 59 Gemma 3 27B 2025-03 Open SourceMultimodalApache 2.0 New Google multimodal 67.9% 78.4% 42% 68.5% N/A 79% N/A N/A 71.5% $0.0003 110 128K 60 GPT-4o mini 2024-07 Cost-Effective OpenAI conversation 67.8% 82% 40.2% 59.4% N/A 87.2% N/A N/A 70.2% $0.007 120 128K 61 Llama 4 Scout 2024-12 Least ExpensiveOpen Source New Meta conversation 67% 74.3% 57.2% 69.4% N/A N/A N/A N/A N/A $0.0003 120 128K 62 Amazon Nova Pro 2025-01 AWS NativeMultimodalCost-Effective New Amazon multimodal 66.9% 80% 44% 63.5% N/A 79% N/A N/A 68% $0.0008 100 300K 63 Claude 3.5 Haiku 2024-11 FastCost-Effective New Anthropic conversation 66% 65% 41.6% N/A N/A 88.1% N/A N/A 69.2% $0.005 140 200K 64 Claude 3 Haiku 2024-03 FastComplete Benchmarks Anthropic conversation 65.3% 75.2% 33.3% 50.2% 85.9% 75.9% 73.7% 88.9% 38.9% $0.004 160 200K 65 Phi-4 Mini 2025-04 Open SourceEdge Deployable3.8B Params New Microsoft conversation 63.4% 75.6% 37.3% N/A N/A 73% N/A N/A 67.5% $0.0001 200 16K 66 GPT-4.1 nano 2024-11 Ultra Fast New OpenAI conversation 61.9% 80.1% 50.3% 55.4% N/A N/A N/A N/A N/A $0.005 150 32K 67 Gemma 3 12B 2025-03 Open SourceLightweightOn-Device New Google conversation 60.2% 74.2% 37% 59.8% N/A 72% N/A N/A 58% $0.0001 160 128K 68 Amazon Nova Lite 2025-01 AWS NativeUltra FastCheapest Multimodal New Amazon conversation 56.6% 73% 33% 55% N/A 68% N/A N/A 54% $0.00006 180 300K 69 o3-pro 2025-01 Upcoming New OpenAI reasoning N/A N/A N/A N/A N/A N/A N/A N/A N/A $0.08 20 128K 70 GPT-4o Realtime 2024-10 RealtimeVoice OpenAI conversation N/A N/A N/A N/A N/A N/A N/A N/A N/A $0.02 200 128K 71 GPT-4o mini Realtime 2024-10 RealtimeVoiceCost-Effective OpenAI conversation N/A N/A N/A N/A N/A N/A N/A N/A N/A $0.008 250 128K ## Want to use AI on your website? Create a free WhisperChat support chatbot. Train it on your website content and start answering repetitive customer questions in minutes. [Start Free](https://www.whisperchat.ai/auth/signup) ## How to Use the LLM Leaderboard Choosing the right AI model matters because cost, speed, and accuracy vary significantly across providers. This leaderboard compares 71+ large language models across eight industry-standard benchmarks so you can make evidence-based decisions. Step 1: Browse the main leaderboard to review overall rankings by composite score. Step 2: Open the Performance Charts tab to visualize strengths across benchmarks like MMLU, GPQA, and HumanEval. Step 3: Use Model Comparison to evaluate 2-3 models side by side on metrics relevant to your use case. Step 4: Review Benchmark Details for scoring methodology and context. Whether you are building customer support automation, coding assistants, or content generation workflows, selecting the right model can save substantial API spend. Use cost and throughput columns to identify your performance-budget sweet spot. A model that scores 5% lower but costs 80% less may be the practical winner. The leaderboard is updated regularly as new models release and benchmarks evolve. Recheck it before major infrastructure decisions so your stack reflects current capabilities and pricing realities. For practical evaluation, shortlist models from the leaderboard and run your own task-specific test set before rollout. Benchmarks provide directional guidance, but domain prompts, latency expectations, and compliance constraints can change final selection. Combining public rankings with internal testing gives the most reliable model choice. ### Related Articles - [What Is Conversational AI?](https://www.whisperchat.ai/blog/what-is-conversational-ai) - [Custom GPT and AI Chatbot Solutions](https://www.whisperchat.ai/blog/custom-gpt-ai-chatbot-solutions) - [AI Tools for Business Automation](https://www.whisperchat.ai/blog/ai-tools-for-business-automation) ### Related Tools - [Chatbot ROI Calculator](https://www.whisperchat.ai/tools/chatbot-roi-calculator) - [AI FAQ Generator](https://www.whisperchat.ai/tools/ai-faq-generator) - [AI Answer Generator](https://www.whisperchat.ai/tools/ai-answer-generator) ## Try our other free tools! Explore more powerful AI tools to enhance your productivity and creativity. GENERATOR ### AI FAQ GENERATOR AI FAQ GENERATOR Generate comprehensive FAQ sections for your website or product using AI. Create helpful answers to common questions automatically. [](https://www.whisperchat.ai/tools/ai-faq-generator) GENERATOR ### AI ANSWER GENERATOR AI ANSWER GENERATOR Create intelligent, contextual answers to any question or query. Perfect for customer support and knowledge base creation. [](https://www.whisperchat.ai/tools/ai-answer-generator) EDITOR ### AI HUMANIZE TEXT AI HUMANIZE TEXT Transform robotic or AI-generated text into natural, human-sounding language. Improve relatability and tone with just one click. [](https://www.whisperchat.ai/tools/ai-humanize-text-generator) GENERATOR ### AI EMAIL RESPONSE GENERATOR AI EMAIL RESPONSE GENERATOR Generate professional email responses tailored to your specific needs. Save time with smart, contextual email automation. [](https://www.whisperchat.ai/tools/ai-email-response-generator) [](https://www.whisperchat.ai/tools)