AI Model Ranking (LLM Leaderboard)
Fastest AI Models
Language models ranked by inference speed and throughput
|
Model
AI model name and provider organization |
Input/1M
Cost per 1 million input tokens (text you send to the model) |
Output/1M
Cost per 1 million output tokens (text the model generates for you) |
MMLU-Pro
Massive Multitask Language Understanding (Professional) - tests broad knowledge across 14 subjects including STEM, humanities, and social sciences |
Speed
Inference throughput in tokens per second - how fast the model generates responses |
GPQA
Graduate-level Google-Proof Q&A benchmark - tests PhD-level reasoning and advanced intelligence |
AIME 2025
American Invitational Mathematics Examination 2025 - tests advanced mathematical problem-solving ability |
Release
When the model was released - newer models may have more capabilities | Compare |
|---|---|---|---|---|---|---|---|---|
 #1 Mercury 2 by Inception | $0.25 | $0.75 | - | 1196 tok/s | 77.0% | - | Feb 20, 2026 | |
 #2 Granite 3.3 8B (Non-reasoning) by IBM | $0.03 | $0.25 | 46.8% | 579 tok/s | 33.8% | 6.7% | Apr 16, 2025 | |
 #3 Gemini 2.5 Flash-Lite Preview (Sep '25) (Reasoning) by Google | $0.10 | $0.40 | 80.8% | 468 tok/s | 70.9% | 68.7% | Sep 8, 2025 | |
 #4 Nova Micro by Amazon | $0.04 | $0.14 | 53.1% | 412 tok/s | 35.8% | 6.0% | Dec 3, 2024 | |
| #5 Granite 4.0 H Small by IBM | $0.06 | $0.25 | 62.4% | 386 tok/s | 41.6% | 13.7% | Sep 22, 2025 | |
| #6 Gemini 2.5 Flash-Lite (Non-reasoning) by Google | $0.10 | $0.40 | 72.4% | 359 tok/s | 47.4% | 35.3% | Jun 17, 2025 | |
| #7 Gemini 2.5 Flash-Lite Preview (Sep '25) (Non-reasoning) by Google | $0.10 | $0.40 | 79.6% | 340 tok/s | 65.1% | 46.7% | Sep 25, 2025 | |
| #8 Gemini 2.5 Flash-Lite (Reasoning) by Google | $0.10 | $0.40 | 75.9% | 329 tok/s | 62.5% | 53.3% | Jun 17, 2025 | |
 #9 gpt-oss-120B (high) by OpenAI | $0.15 | $0.60 | 80.8% | 327 tok/s | 78.2% | 93.4% | Aug 5, 2025 | |
| #10 gpt-oss-120B (low) by OpenAI | $0.15 | $0.59 | 77.5% | 321 tok/s | 67.2% | 66.7% | Aug 5, 2025 | |
| #11 gpt-oss-20B (high) by OpenAI | $0.07 | $0.23 | 74.8% | 312 tok/s | 68.8% | 89.3% | Aug 5, 2025 | |
| #12 GPT-5 Codex (high) by OpenAI | $1.25 | $10.00 | 86.5% | 310 tok/s | 83.7% | 98.7% | Sep 23, 2025 | |
| #13 gpt-oss-20B (low) by OpenAI | $0.07 | $0.20 | 71.8% | 307 tok/s | 61.1% | 62.3% | Aug 5, 2025 | |
 #14 Ministral 3 3B by Mistral | $0.10 | $0.10 | 52.4% | 295 tok/s | 35.8% | 22.0% | Dec 2, 2025 | |
| #15 Gemini 2.5 Flash (Reasoning) by Google | $0.30 | $2.50 | 83.2% | 285 tok/s | 79.0% | 73.3% | May 20, 2025 | |
 #16 Grok Code Fast 1 by xAI | $0.20 | $1.50 | 79.3% | 278 tok/s | 72.7% | 43.3% | Aug 28, 2025 | |
| #17 Gemini 2.5 Flash (Non-reasoning) by Google | $0.30 | $2.50 | 80.9% | 253 tok/s | 68.3% | 60.3% | May 20, 2025 | |
| #18 Nova 2.0 Lite (medium) by Amazon | $0.30 | $2.50 | 81.3% | 243 tok/s | 76.8% | 88.7% | Oct 29, 2025 | |
| #19 Devstral Small (Jul '25) by Mistral | $0.10 | $0.30 | 62.2% | 234 tok/s | 41.4% | 29.3% | Jul 10, 2025 | |
| #20 Mistral Small 3 by Mistral | $0.10 | $0.30 | 65.2% | 231 tok/s | 46.2% | 4.3% | Jan 30, 2025 | |
| #21 Nova Lite by Amazon | $0.06 | $0.24 | 59.0% | 228 tok/s | 43.3% | 7.0% | Dec 3, 2024 | |
| #22 Nova 2.0 Lite (low) by Amazon | $0.30 | $2.50 | 78.8% | 219 tok/s | 69.8% | 46.7% | Oct 29, 2025 | |
| #23 Magistral Small 1.2 by Mistral | $0.50 | $1.50 | 76.8% | 213 tok/s | 66.3% | 80.3% | Sep 17, 2025 | |
| #24 GPT-5.1 Codex (high) by OpenAI | $1.25 | $10.00 | 86.0% | 210 tok/s | 86.0% | 95.7% | Nov 13, 2025 | |
| #25 Gemini 3 Flash Preview (Reasoning) by Google | $0.50 | $3.00 | 89.0% | 208 tok/s | 89.8% | 97.0% | Dec 17, 2025 | |
 #26 GLM-4.5-Air by Z AI | $0.20 | $1.10 | 81.5% | 208 tok/s | 73.3% | 80.7% | Jul 28, 2025 | |
| #27 Nova 2.0 Omni (Non-reasoning) by Amazon | $0.30 | $2.50 | 71.9% | 207 tok/s | 55.5% | 37.0% | Nov 26, 2025 | |
| #28 GPT-5 (ChatGPT) by OpenAI | $1.25 | $10.00 | 82.0% | 202 tok/s | 68.6% | 48.3% | Aug 7, 2025 | |
 #29 Qwen3 0.6B (Reasoning) by Alibaba | $0.11 | $1.26 | 34.7% | 201 tok/s | 23.9% | 18.0% | Apr 28, 2025 | |
| #30 Devstral Small 2 by Mistral | N/A | N/A | 67.8% | 201 tok/s | 53.2% | 34.3% | Dec 9, 2025 | |
| #31 Grok 3 mini Reasoning (high) by xAI | $0.30 | $0.50 | 82.8% | 193 tok/s | 79.1% | 84.7% | Feb 19, 2025 | |
| #32 Nova 2.0 Lite (Non-reasoning) by Amazon | $0.30 | $2.50 | 74.3% | 192 tok/s | 60.3% | 33.7% | Oct 29, 2025 | |
| #33 Qwen3 0.6B (Non-reasoning) by Alibaba | $0.11 | $0.42 | 23.1% | 189 tok/s | 23.1% | 10.3% | Apr 28, 2025 | |
 #34 Jamba 1.6 Mini by AI21 Labs | $0.20 | $0.40 | 36.7% | 180 tok/s | 30.0% | - | Mar 6, 2025 | |
 #35 Llama 3.1 Instruct 8B by Meta | $0.10 | $0.10 | 47.6% | 180 tok/s | 25.9% | 4.3% | Jul 23, 2024 | |
| #36 Gemini 3 Flash Preview (Non-reasoning) by Google | $0.50 | $3.00 | 88.2% | 176 tok/s | 81.2% | 55.7% | Dec 17, 2025 | |
 #37 MiMo-V2-Flash (Reasoning) by Xiaomi | $0.10 | $0.30 | 84.3% | 173 tok/s | 84.6% | 96.3% | Dec 16, 2025 | |
| #38 Grok 4 Fast (Reasoning) by xAI | $0.20 | $0.50 | 85.0% | 172 tok/s | 84.7% | 89.7% | Sep 19, 2025 | |
 #39 Olmo 3 7B Think by Allen Institute for AI | $0.12 | $0.20 | 65.5% | 171 tok/s | 51.6% | 70.7% | Nov 20, 2025 | |
| #40 Ministral 3 8B by Mistral | $0.15 | $0.15 | 64.2% | 169 tok/s | 47.1% | 31.7% | Dec 2, 2025 | |
| #41 MiMo-V2-Flash (Feb 2026) by Xiaomi | $0.10 | $0.30 | - | 167 tok/s | 83.5% | - | Dec 16, 2025 | |
| #42 GPT-5.1 Codex mini (high) by OpenAI | $0.25 | $2.00 | 82.0% | 162 tok/s | 81.3% | 91.7% | Nov 13, 2025 | |
| #43 Llama 4 Scout by Meta | $0.18 | $0.66 | 75.2% | 157 tok/s | 58.7% | 14.0% | Apr 5, 2025 | |
| #44 Grok 4.1 Fast (Reasoning) by xAI | $0.20 | $0.50 | 85.4% | 156 tok/s | 85.3% | 89.3% | Nov 19, 2025 | |
| #45 Nova 2.0 Pro Preview (Non-reasoning) by Amazon | $1.25 | $10.00 | 77.2% | 156 tok/s | 63.6% | 30.7% | Nov 27, 2025 | |
| #46 Mistral 7B Instruct by Mistral | $0.25 | $0.25 | 24.5% | 153 tok/s | 17.7% | - | Sep 27, 2023 | |
| #47 MiMo-V2-Flash (Non-reasoning) by Xiaomi | $0.10 | $0.30 | 74.4% | 151 tok/s | 65.6% | 67.7% | Dec 16, 2025 | |
| #48 GPT-4o (Nov '24) by OpenAI | $2.50 | $10.00 | 74.8% | 151 tok/s | 54.3% | 6.0% | Nov 20, 2024 | |
| #49 Gemini 2.5 Pro by Google | $1.25 | $10.00 | 86.2% | 150 tok/s | 84.4% | 87.7% | Jun 5, 2025 | |
| #50 GLM-4.7 (Reasoning) by Z AI | $0.55 | $2.15 | 85.6% | 149 tok/s | 85.9% | 95.0% | Dec 22, 2025 | |
| #51 o3-mini (high) by OpenAI | $1.10 | $4.40 | 80.2% | 149 tok/s | 77.3% | - | Jan 31, 2025 | |
| #52 Grok 4 Fast (Non-reasoning) by xAI | $0.20 | $0.50 | 73.0% | 149 tok/s | 60.6% | 41.3% | Sep 19, 2025 | |
| #53 Qwen3 30B A3B 2507 (Reasoning) by Alibaba | $0.20 | $2.40 | 80.5% | 147 tok/s | 70.7% | 56.3% | Jul 30, 2025 | |
 #54 Apriel-v1.5-15B-Thinker by ServiceNow | N/A | N/A | 77.3% | 145 tok/s | 71.3% | 87.5% | Sep 30, 2025 | |
| #55 o1 by OpenAI | $15.00 | $60.00 | 84.1% | 145 tok/s | 74.7% | - | Dec 5, 2024 | |
 #56 Ling-mini-2.0 by InclusionAI | $0.07 | $0.28 | 67.1% | 143 tok/s | 56.2% | 49.3% | Sep 9, 2025 | |
| #57 Gemini 3 Pro Preview (low) by Google | $2.00 | $12.00 | 89.5% | 142 tok/s | 88.7% | 86.7% | Nov 18, 2025 | |
| #58 Qwen3 Next 80B A3B (Reasoning) by Alibaba | $0.50 | $6.00 | 82.4% | 138 tok/s | 75.9% | 84.3% | Sep 11, 2025 | |
| #59 Apriel-v1.6-15B-Thinker by ServiceNow | N/A | N/A | 79.0% | 138 tok/s | 73.3% | 88.0% | Nov 25, 2025 | |
| #60 Llama 3.2 Instruct 1B by Meta | $0.10 | $0.10 | 20.0% | 136 tok/s | 19.6% | - | Sep 25, 2024 | |
 #61 NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning) by NVIDIA | $0.20 | $0.60 | 64.9% | 136 tok/s | 43.9% | 26.7% | Oct 28, 2025 | |
| #62 Mistral Small 3.2 by Mistral | $0.10 | $0.30 | 68.1% | 135 tok/s | 50.5% | 27.0% | Jun 20, 2025 | |
| #63 o3-mini by OpenAI | $1.10 | $4.40 | 79.1% | 135 tok/s | 74.8% | - | Jan 31, 2025 | |
| #64 Nova 2.0 Pro Preview (medium) by Amazon | $1.25 | $10.00 | 83.0% | 133 tok/s | 78.5% | 89.0% | Nov 27, 2025 | |
| #65 Qwen3 Next 80B A3B Instruct by Alibaba | $0.50 | $2.00 | 81.9% | 133 tok/s | 73.8% | 66.3% | Sep 11, 2025 | |
| #66 NVIDIA Nemotron Nano 12B v2 VL (Reasoning) by NVIDIA | $0.20 | $0.60 | 75.9% | 132 tok/s | 57.2% | 75.0% | Oct 28, 2025 | |
| #67 Nova 2.0 Pro Preview (low) by Amazon | $1.25 | $10.00 | 82.2% | 131 tok/s | 75.1% | 63.3% | Nov 27, 2025 | |
| #68 Gemini 3 Pro Preview (high) by Google | $2.00 | $12.00 | 89.8% | 131 tok/s | 90.8% | 95.7% | Nov 18, 2025 | |
| #69 GPT-5 nano (high) by OpenAI | $0.05 | $0.40 | 78.0% | 130 tok/s | 67.6% | 83.7% | Aug 7, 2025 | |
 #70 Sonar by Perplexity | $1.00 | $1.00 | 68.9% | 129 tok/s | 47.1% | - | Jan 21, 2025 | |
| #71 GLM-4.7 (Non-reasoning) by Z AI | $0.55 | $2.15 | 79.4% | 129 tok/s | 66.4% | 48.0% | Dec 22, 2025 | |
| #72 GPT-5 nano (medium) by OpenAI | $0.05 | $0.40 | 77.2% | 128 tok/s | 67.0% | 78.3% | Aug 7, 2025 | |
| #73 NVIDIA Nemotron 3 Nano 30B A3B (Reasoning) by NVIDIA | $0.06 | $0.24 | 79.4% | 128 tok/s | 75.7% | 91.0% | Dec 15, 2025 | |
| #74 Ministral 3 14B by Mistral | $0.20 | $0.20 | 69.3% | 124 tok/s | 57.2% | 30.0% | Dec 2, 2025 | |
| #75 GPT-5.1 (high) by OpenAI | $1.25 | $10.00 | 87.0% | 124 tok/s | 87.3% | 94.0% | Nov 13, 2025 | |
| #76 Qwen3 1.7B (Reasoning) by Alibaba | $0.11 | $1.26 | 57.0% | 124 tok/s | 35.6% | 38.7% | Apr 28, 2025 | |
| #77 NVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning) by NVIDIA | $0.05 | $0.20 | 57.9% | 123 tok/s | 39.9% | 13.3% | Dec 15, 2025 | |
| #78 Qwen3 VL 8B (Reasoning) by Alibaba | $0.18 | $2.10 | 74.9% | 123 tok/s | 57.9% | 30.7% | Oct 14, 2025 | |
| #79 o4-mini (high) by OpenAI | $1.10 | $4.40 | 83.2% | 123 tok/s | 78.4% | 90.7% | Apr 16, 2025 | |
 #80 Claude 4.5 Haiku (Reasoning) by Anthropic | $1.00 | $5.00 | 76.0% | 121 tok/s | 67.2% | 83.7% | Oct 15, 2025 | |
| #81 Claude 3 Haiku by Anthropic | $0.25 | $1.25 | - | 120 tok/s | 37.4% | - | Mar 4, 2024 | |
| #82 o3 by OpenAI | $2.00 | $8.00 | 85.3% | 120 tok/s | 82.7% | 88.3% | Apr 16, 2025 | |
| #83 GPT-4o (May '24) by OpenAI | $5.00 | $15.00 | 74.0% | 118 tok/s | 52.6% | - | May 13, 2024 | |
| #84 Qwen3 VL 8B Instruct by Alibaba | $0.18 | $0.70 | 68.6% | 117 tok/s | 42.7% | 27.3% | Oct 14, 2025 | |
| #85 NVIDIA Nemotron Nano 9B V2 (Non-reasoning) by NVIDIA | $0.06 | $0.23 | 73.9% | 115 tok/s | 55.7% | 62.3% | Aug 18, 2025 | |
| #86 GPT-5 nano (minimal) by OpenAI | $0.05 | $0.40 | 55.6% | 114 tok/s | 42.8% | 27.3% | Aug 7, 2025 | |
| #87 Mistral Small (Feb '24) by Mistral | $1.00 | $3.00 | 41.9% | 114 tok/s | 30.2% | - | Feb 26, 2024 | |
| #88 Molmo2-8B by Allen Institute for AI | N/A | N/A | - | 114 tok/s | 42.5% | - | Dec 11, 2025 | |
| #89 Qwen3 1.7B (Non-reasoning) by Alibaba | $0.11 | $0.42 | 41.1% | 114 tok/s | 28.3% | 7.3% | Apr 28, 2025 | |
| #90 GPT-4.1 nano by OpenAI | $0.10 | $0.40 | 65.7% | 114 tok/s | 51.2% | 24.0% | Apr 14, 2025 | |
| #91 Llama 4 Maverick by Meta | $0.31 | $0.85 | 80.9% | 112 tok/s | 67.1% | 19.3% | Apr 5, 2025 | |
| #92 Devstral Medium by Mistral | $0.40 | $2.00 | 70.8% | 112 tok/s | 49.2% | 4.7% | Jul 10, 2025 | |
| #93 Mistral Small (Sep '24) by Mistral | $0.20 | $0.60 | 52.9% | 111 tok/s | 38.1% | - | Sep 17, 2024 | |
| #94 Llama 2 Chat 7B by Meta | $0.05 | $0.25 | 16.4% | 110 tok/s | 22.7% | - | Jul 18, 2023 | |
| #95 Grok 4.1 Fast (Non-reasoning) by xAI | $0.20 | $0.50 | 74.3% | 110 tok/s | 63.7% | 34.3% | Nov 19, 2025 | |
| #96 Claude 4.5 Haiku (Non-reasoning) by Anthropic | $1.00 | $5.00 | 80.0% | 110 tok/s | 64.6% | 39.0% | Oct 15, 2025 | |
| #97 Sonar Pro by Perplexity | $3.00 | $15.00 | 75.5% | 109 tok/s | 57.8% | - | Jan 21, 2025 | |
| #98 Mistral Small 3.1 by Mistral | $0.10 | $0.30 | 65.9% | 107 tok/s | 45.4% | 3.7% | Mar 17, 2025 | |
| #99 NVIDIA Nemotron Nano 9B V2 (Reasoning) by NVIDIA | $0.04 | $0.16 | 74.2% | 106 tok/s | 57.0% | 69.7% | Aug 18, 2025 | |
| #100 Qwen3 Coder Next by Alibaba | $0.20 | $1.20 | - | 105 tok/s | 73.7% | - | Feb 3, 2026 |
Made in Europe Chat with 100+ AI Models in one App.
Use Claude, ChatGPT, Gemini alongside with EU-Hosted Models like Deepseek, GLM-5, Kimi K2.5 and many more.
Understanding the AI Model Leaderboard
This comprehensive AI model leaderboard helps you compare and choose the best large language models (LLMs) for your needs. We track standardized AI benchmarks, token pricing, inference speed, and model capabilities across all major AI providers like OpenAI, Anthropic, Google, Meta, and DeepSeek.
Core AI Benchmarks Explained
- MMLU-Pro: Tests broad knowledge across 14 academic subjects including STEM, humanities, and social sciences - the foundational intelligence benchmark
- GPQA: Graduate-level Google-Proof Q&A benchmark - measures PhD-level reasoning and advanced problem-solving capabilities
- AIME 2025: American Invitational Mathematics Examination - evaluates elite mathematical reasoning and competition-level problem solving
- Coding Index: Composite score of LiveCodeBench, SciCode, and coding benchmarks - measures programming ability
- Math Index: Composite score of AIME, MATH-500, and mathematical reasoning tests
Key Metrics to Consider
- Token Pricing: Compare input vs output token costs per million - crucial for estimating API expenses and optimizing usage patterns
- Inference Speed: Measured in tokens/second - determines response time for chatbots, streaming, and real-time applications
- Release Date: Newer models often incorporate latest training techniques and updated knowledge cutoffs
- Benchmark Scores: Percentage scores (0-100%) make it easy to compare model capabilities at a glance
How to Choose the Right AI Model for Your Use Case
For Research & Analysis
Prioritize models with high MMLU-Pro (70%+) and GPQA (60%+) scores for complex reasoning tasks, academic research, and technical documentation
For Cost Optimization
Sort by input/output pricing - smaller models often deliver 80% of flagship performance at 10% of the cost for simple tasks
For Math & STEM
Filter by Math Index or AIME 2025 scores (50%+) for quantitative analysis, engineering calculations, and scientific applications
All benchmark scores and pricing data are updated daily from Artificial Analysis to reflect the latest model versions and capabilities. Use the sort filters above to find AI models by intelligence, cost, coding ability, math performance, speed, or release date.
Frequently Asked Questions
What is MMLU-Pro and why is it the standard AI intelligence benchmark?
MMLU-Pro (Massive Multitask Language Understanding - Professional) is the most comprehensive AI benchmark, testing models across 14 academic subjects including mathematics, science, history, law, and ethics. Scores range from 46% (basic competency) to 87% (near-expert level). Models scoring above 75% demonstrate strong general intelligence suitable for professional applications, while scores below 60% indicate limitations in complex reasoning tasks.
What does GPQA measure and which models score highest?
GPQA (Graduate-level Google-Proof Q&A) tests PhD-level reasoning with questions designed to be "Google-proof" - requiring deep understanding rather than simple fact retrieval. Top models like GPT-5.1 (87.3%), GPT-5 mini (82.8%), and o3 (82.7%) excel at GPQA, making them ideal for research, technical analysis, and complex problem-solving. Models below 50% GPQA struggle with advanced reasoning and may provide superficial answers to complex questions.
What is AIME 2025 and how does it evaluate AI mathematical ability?
AIME 2025 (American Invitational Mathematics Examination) is an elite math competition benchmark that tests advanced problem-solving, algebra, geometry, and number theory. Scores above 80% (like GPT-5 Codex at 98.7% or GPT-5.1 at 94%) indicate exceptional mathematical reasoning suitable for engineering, scientific computing, and quantitative analysis. Models scoring below 50% may struggle with multi-step mathematical problems or require explicit problem breakdown.
How is AI model pricing calculated and what's considered cost-effective?
AI model pricing is measured per 1 million tokens (approximately 750,000 words). Input pricing covers text you send, while output pricing covers generated responses. Budget models like Llama 3.3 70B cost $0.54/$0.71 per million tokens, mid-tier models like GPT-5 nano cost $0.05/$0.40, while premium models like GPT-5 cost $1.25/$10. For typical applications with 3:1 input-to-output ratio, budget models can be 10-20x cheaper than flagship models while maintaining 70-80% performance.
Which AI models are best for coding and programming tasks?
Sort by Coding Index to see top programming models. Our Coding Index combines LiveCodeBench, SciCode, and coding benchmarks. Top performers include GPT-5.1 (57.5 index), GPT-5 mini (51.4), and GPT-5 Codex (53.5). These models excel at code generation, debugging, refactoring, and explaining complex algorithms. For budget-conscious developers, models with 40+ coding index scores offer excellent value for routine programming tasks.
How often are AI model benchmarks and rankings updated?
Our leaderboard syncs daily with Artificial Analysis API to ensure benchmark scores (MMLU-Pro, GPQA, AIME 2025), pricing, and inference speed data reflect the latest model versions. New model releases appear immediately under the "Newest" sort option. Benchmark scores can change when providers release updated versions - for example, GPT-5.1 released in November 2025 achieved 69.7 intelligence compared to GPT-5's 68.5 from August 2025.
What inference speed (tokens/second) do I need for my application?
Inference speed determines how fast models generate responses. For real-time chatbots and interactive applications, target 100+ tokens/second (models like gpt-oss-120B at 340 tok/s). For background processing and batch jobs, 50-100 tok/s is sufficient. Premium reasoning models like GPT-5 (103 tok/s) balance speed and capability. Note that higher inference speed doesn't always mean better quality - slower models often deliver more thoughtful, detailed responses.
Can I test these AI models for free before committing?
Yes! Try our free AI chat interface to test different models instantly without creating an account. Many providers also offer free tiers: OpenAI (ChatGPT with daily limits), Anthropic (Claude with usage caps), Google (Gemini free tier), and open-source models like Llama 3.3. Compare performance on your specific use case before upgrading to paid plans.