LLM Leaderboard - Comparison of AI models from OpenAI, Anthropic, Google, Meta & others
Comparison and ranking the performance of over 250 AI models (LLMs) across key metrics including intelligence, price, performance and speed (time per task in minutes & latency as time to first answer token), context window & others.
For more details including relating to our methodology, see our FAQs.
Intelligence
Claude Fable 5.2 (max) and Claude Fable 5.2 (xhigh) are the highest intelligence models, followed by Claude Fable 5.2 (high) and Gemini 4 Carbon (high).
Time per Task
GPT-6 Luna (low) and GPT-6.1 Terra (low) have the lowest time per task, followed by Gemini 4 Hydrogen (low) and Gemini 3.5 Flash-Lite.
Time to First Answer Token
Tiny Aya Global and Phi-4 Multimodal have the lowest time to first answer token, followed by Celeris-1 and NVIDIA Nemotron Nano 9B V2.
Cost per Task
K2 Horizon 375B A23B and Command A+ have the lowest cost per task, followed by North Mini Code and GPT-6 Luna (low).
Context Window
Llama 4 Scout and GPT-6 Sol (Daybreak Blue, max) support the largest context windows, followed by Muse Wonder (max) and Muse Wonder (high).
Further Analysis | ||||||
|---|---|---|---|---|---|---|
65 | $26.54 | 45.3 | 268.43 | 1M | ||
60 | $5.48 | 31.3 | 69.56 | 1M | ||
59 | $10.04 | 25.4 | 399.36 | 1M | ||
57 | $6.24 | 23.8 | 251.67 | 1M | ||
55 | $1.27 | 23.1 | 200.68 | 1M | ||
55 | $10.02 | 33.9 | 278.78 | 1M | ||
54 | $3.16 | 14.8 | 43.34 | 1M | ||
51 | $0.96 | 14.6 | 184.80 | 1M | ||
50 | $4.97 | 29.4 | 45.20 | 1M | ||
| Muse Wonder Full Max | 50 | -- | -- | -- | 1M | |
 Kimi K3.1 Pro | 46 | $3.68 | 39.7 | 185.95 | 1M | |
44 | $6.95 | 14.6 | 26.13 | 1M | ||
36 | $0.41 | 14.9 | 176.72 | 1M | ||
36 | $0.14 | 3.3 | 4.98 | 1M | ||
35 | $7.77 | 26.9 | 62.70 | 500k | ||
 Tencent Hy4 | 35 | $1.03 | 14.2 | 109.88 | 1M | |
 Qwen3.8 Max (0902) | 34 | $10.02 | 48.6 | 114.22 | 984k | |
33 | $0.60 | 14.0 | 45.09 | 1M | ||
32 | $5.01 | 25.8 | 62.23 | 1M | ||
31 | $7.77 | 14.4 | 17.34 | 1M | ||
 Inkling 2 Small | 31 | $1.18 | 31.9 | 69.01 | 1M | |
 MiMo-V2.6-Pro | 30 | $0.26 | 35.6 | 38.58 | 1M | |
 Ling 3.1 Flash 1019 | 30 | $3.19 | -- | -- | 1M | |
 MiniMax-M3.1-Flash | 29 | $0.49 | 16.8 | 59.89 | 1M | |
| Qwen3.8-Flash-Next | 29 | $0.57 | 34.7 | 96.11 | 256k | |
 Step 5 Preview | 29 | $2.35 | 14.5 | 24.28 | 1M | |
28 | $0.15 | 12.0 | 89.04 | 1M | ||
 GLM-5.3-Flash | 28 | $0.58 | 46.8 | 104.76 | 1M | |
| MiMo-V2.6-Flash | 27 | $0.13 | 50.5 | 31.97 | 1M | |
26 | $2.06 | 13.1 | 41.27 | 1M | ||
| Muse Spark Open Xhigh | 25 | $2.02 | 7.9 | 12.45 | 1M | |
 Agnes 2.5 Pro Theta | 24 | $0.57 | 25.5 | 6.33 | 1M | |
23 | $1.77 | 19.7 | 55.72 | 1M | ||
 Nanbeige4.3-33B | 20 | -- | -- | -- | 524k | |
20 | $3.53 | 43.7 | 138.53 | 256k | ||
 JT-4.1 Flash 236B A21B | 19 | -- | -- | -- | 262k | |
| MiniMax-M3 | 19 | $1.29 | 13.7 | 80.76 | 1M | |
| K2 Horizon 375B A23B | 18 | $0.00 | 4.3 | 15.93 | 524k | |
 Beam 1.0 | 18 | $0.28 | 2.1 | 6.67 | 1M | |
 Solar Pro 4 | 14 | $0.96 | 62.5 | 35.11 | 512k | |
 Apodex 1.1 | 14 | $0.76 | 14.8 | 15.77 | 256k | |
| Nemotron 3.5 Super VL | 14 | $0.56 | 6.1 | 15.93 | 1M | |
| Inkling Small | 13 | $0.20 | 5.8 | 19.53 | 1M | |
| Inkling Xhigh | 12 | $1.35 | 11.0 | 21.88 | 1M | |
| Nemotron 3 Ultra 550B A55B Reasoning | 12 | $2.21 | 4.3 | 4.68 | 262k | |
| Gemini 3.5 Flash-Lite | 12 | $0.69 | 1.7 | 4.63 | 1M | |
 K-EXAONE 2.0 | 8 | -- | -- | -- | 262k | |
| Muse Glimmer High | 6 | $0.18 | 5.4 | 25.30 | 131k | |
| Command A+ | 6 | $0.00 | 3.1 | 17.45 | 192k | |
| Nemotron 3.5 Lightning | 6 | $0.31 | 4.2 | 14.20 | 1M | |
 Granite 4.2 8B | 5 | $0.0405 | 11.5 | 54.00 | 131k | |
| Granite 4.2 3B | 5 | -- | -- | -- | 131k | |
 Mistral Medium 3.5 | 5 | $1.04 | 4.5 | 23.58 | 256k | |
4 | -- | -- | 26.30 | 256k | ||
4 | $0.0399 | 1.7 | 10.62 | 256k | ||
4 | $0.35 | 7.5 | 14.96 | 131k | ||
| North Mini Code | 3 | $0.00 | 2.9 | 4.36 | 256k | |
| NVIDIA Nemotron 3 Nano 4B | 2 | -- | -- | -- | 262k | |
| JT-35B-Flash | -- | -- | -- | -- | 256k | |
| Qwen3.5 Omni Plus | -- | -- | -- | 2.83 | 256k | |
-- | -- | -- | -- | 1M | ||
-- | -- | -- | 4.42 | 262k | ||
-- | -- | -- | -- | 262k | ||
| Gemma 3 270M | -- | -- | -- | -- | 32k | |
-- | -- | -- | -- | 262k | ||
| Qwen3 Omni 30B A3B Instruct | -- | -- | -- | 1.14 | 66k | |
-- | -- | -- | 17.15 | 256k | ||
-- | -- | -- | 41.65 | 262k | ||
| Granite 4.2 30B | -- | -- | -- | -- | 131k | |
| K2 Horizon 7B | -- | -- | -- | -- | 524k | |
-- | -- | -- | 2.18 | 131k | ||
| o3 | -- | -- | -- | 7.21 | 200k | |
| Granite 4.0 Micro | -- | -- | -- | -- | 128k | |
-- | -- | -- | 1.06 | 262k | ||
-- | -- | -- | -- | 128k | ||
-- | -- | -- | 0.87 | 1M | ||
| Qwen3 Next 80B A3B Instruct | -- | -- | -- | 1.34 | 262k | |
| Phi-4 Multimodal Instruct | -- | -- | -- | 0.38 | 128k | |
-- | -- | -- | 1.22 | 131k | ||
-- | -- | -- | -- | 131k | ||
 G9v3-3B | -- | -- | -- | -- | 131k | |
-- | -- | -- | 1.38 | 262k | ||
| K2 Horizon 3.7B | -- | -- | -- | -- | 524k | |
 Olmo 3 7B Think | -- | -- | -- | -- | 66k | |
 HyperCLOVA X SEED Think (32B) | -- | -- | -- | -- | 128k | |
 HyperNova 60B 2605 High, Based on gpt-oss-120b | -- | -- | -- | -- | 131k | |
 Nova Micro | -- | -- | -- | 1.38 | 130k | |
-- | -- | -- | -- | 262k | ||
 INTELLECT-3 Based on GLM-4.5-Air | -- | -- | -- | -- | 131k | |
| Olmo 3.1 32B Think | -- | -- | -- | -- | 66k | |
 Nex-N2-Pro Based on Qwen3.5-397B-A17B | -- | -- | -- | -- | 262k | |
| Command A | -- | -- | -- | 10.19 | 256k | |
 LFM2.5-VL-1.6B | -- | -- | -- | -- | 32k | |
| JT-MINI | -- | -- | -- | -- | 128k | |
| Molmo2-8B | -- | -- | -- | -- | 37k | |
-- | -- | -- | -- | 262k | ||
| Ling-3.0-flash-VL | -- | -- | -- | -- | 262k | |
 MiniCPM-V 4.6 1.3B | -- | -- | -- | -- | 262k | |
 Sarvam 105B High | -- | -- | -- | 10.52 | 128k | |
| Llama 3.1 Instruct 405B | -- | -- | -- | -- | 128k | |
| Ring-flash-2.0 | -- | -- | -- | -- | 128k | |
| LFM2.5-1.2B-Instruct | -- | -- | -- | -- | 32k | |
 Apertus 70B Instruct | -- | -- | -- | -- | 66k | |
 Doubao Seed Code | -- | -- | -- | -- | 256k | |
-- | -- | -- | 1.56 | 128k | ||
| K2 Horizon MoVA 36B A4B | -- | -- | -- | -- | 524k | |
| Granite 4.0 350M | -- | -- | -- | -- | 33k | |
| Kimi Linear 48B A3B Instruct | -- | -- | -- | -- | 1M | |
| Solar Open 100B Reasoning | -- | -- | -- | -- | 128k | |
| GPT-5.3 Codex Xhigh | -- | -- | -- | 48.50 | 400k | |
| LFM2.5-1.2B-Thinking | -- | -- | -- | -- | 32k | |
| K2 Think V2 | -- | -- | -- | -- | 262k | |
 LongCat 2.0 | -- | -- | -- | 46.21 | 1M | |
 Tri-21B-Think | -- | -- | -- | -- | 32k | |
| Granite 4.0 H Small | -- | -- | -- | -- | 128k | |
| Llama 3.2 Instruct 11B (Vision) | -- | -- | -- | 4.36 | 128k | |
| LongCat Flash Lite | -- | -- | -- | -- | 256k | |
| Nemotron 3.5 Bolt E4B 4GB Reasoning | -- | -- | -- | -- | 131k | |
-- | -- | -- | 19.39 | 1M | ||
| Solar Mini 4 | -- | -- | -- | 98.60 | 1M | |
| MiniCPM5-2B | -- | -- | -- | -- | 131k | |
 Motif-2-12.7B-Reasoning | -- | -- | -- | -- | 128k | |
-- | -- | -- | 3.15 | 262k | ||
| Ministral 3 3B | -- | -- | -- | 1.50 | 256k | |
| Ling-3.0-flash-Fin | -- | -- | -- | -- | 262k | |
-- | -- | -- | 27.97 | 256k | ||
-- | -- | -- | -- | 128k | ||
| Qwen3.5 Omni Flash | -- | -- | -- | 2.85 | 256k | |
| Magistral Small 1.2 | -- | -- | -- | 3.26 | 128k | |
| Ministral 3 14B | -- | -- | -- | 2.69 | 256k | |
-- | -- | -- | -- | 128k | ||
| Sarvam 30B High | -- | -- | -- | -- | 66k | |
 Mercury 2.5 | -- | -- | -- | 2.30 | 260k | |
 DeepSeek V4 Flash Vision Max | -- | -- | -- | 45.55 | 1M | |
-- | -- | -- | -- | 128k | ||
 Jamba 1.7 Large | -- | -- | -- | -- | 256k | |
| Nemotron 3.5 Bolt 74B-A9B | -- | -- | -- | -- | 262k | |
| Nanbeige4.1-3B | -- | -- | -- | -- | 256k | |
| Solar Open2 250B | -- | -- | -- | -- | 1M | |
| Ling 3.0 Tiny | -- | -- | -- | 11.57 | 262k | |
| Celeris-1 | -- | -- | -- | 0.69 | 131k | |
| Tiny Aya Global | -- | -- | -- | 0.20 | 8k | |
 A.X-K2 | -- | -- | -- | -- | 262k | |
-- | -- | -- | -- | 64k | ||
-- | -- | -- | 18.78 | 128k | ||
| Magistral Medium 1.2 | -- | -- | -- | -- | 128k | |
| Apertus 8B Instruct | -- | -- | -- | -- | 66k | |
| Nemotron 3 Nano Omni 30B A3B Reasoning | -- | -- | -- | -- | 256k | |
| Kimi K2.7 Code | -- | -- | -- | 43.69 | 256k | |
| Qwen3.8 2.4T A95B | -- | -- | -- | 112.78 | 262k | |
| Olmo 3 7B Instruct | -- | -- | -- | -- | 66k | |
| Llama 4 Maverick | -- | -- | -- | 1.64 | 1M | |
| GPT-6 Sol Daybreak Blue, max | -- | -- | -- | -- | 1M | |
| Olmo 3.1 32B Instruct | -- | -- | -- | -- | 66k | |
| Ministral 3 8B | -- | -- | -- | 1.85 | 256k | |
| Llama 3.3 Instruct 70B | -- | -- | -- | 1.71 | 128k | |
| Nemotron Cascade 2 30B A3B | -- | -- | -- | -- | 1M | |
 R1 1776 | -- | -- | -- | -- | 128k | |
| Llama 3.2 Instruct 90B (Vision) | -- | -- | -- | -- | 128k | |
| K2 Horizon 0.9B | -- | -- | -- | -- | 131k | |
| Phi-4 Mini Instruct | -- | -- | -- | 4.57 | 128k | |
-- | -- | -- | -- | 512k | ||
| Gemini 3 Deep Think | -- | -- | -- | -- | 128k | |
 ERNIE 5.0 Thinking Preview | -- | -- | -- | -- | 128k | |
| Ling 3.1 Flash | -- | -- | -- | 44.22 | 1M | |
| Granite 4.0 H 350M | -- | -- | -- | -- | 33k | |
 Apriel-v1.6-15B-Thinker | -- | -- | -- | -- | 128k | |
| Qwen3.7 Plus | -- | -- | -- | 49.15 | 1M | |
 DeepHermes 3 - Mistral 24B Preview Non-reasoning | -- | -- | -- | -- | 32k | |
| Step3 VL 10B | -- | -- | -- | -- | 66k | |
| Llama 4 Scout | -- | -- | -- | 1.13 | 10M | |
| ERNIE 4.5 300B A47B | -- | -- | -- | -- | 131k | |
| Molmo 7B-D | -- | -- | -- | -- | 4k | |
| Motif 3 | -- | -- | -- | -- | 262k | |
| Hy3 | -- | -- | -- | 44.55 | 256k | |
| Jamba 1.7 Mini | -- | -- | -- | -- | 258k | |
 Trinity Large Thinking | -- | -- | -- | -- | 512k | |
| DeepHermes 3 - Llama-3.1 8B Preview Non-reasoning | -- | -- | -- | -- | 128k | |
| Qwen3 Omni 30B A3B (Reasoning) | -- | -- | -- | 25.55 | 66k | |
| Gemini 3.1 Pro Preview | -- | -- | -- | 15.28 | 1M | |
| G9v3-39A5B | -- | -- | -- | -- | 131k | |
| DiffusionGemma 26B A4B | -- | -- | -- | -- | 256k | |
| GPT-5.5 Instant (June 2026) | -- | -- | -- | 1.87 | 400k | |
 Falcon-H1R-7B | -- | -- | -- | -- | 256k | |
| Qwen3 Next 80B A3B (Reasoning) | -- | -- | -- | -- | 262k | |
| Llama 3.1 Nemotron Ultra 253B v1 Reasoning | -- | -- | -- | -- | 128k | |
| Granite 4.0 H 1B | -- | -- | -- | -- | 128k | |
 Reka Flash 3 | -- | -- | -- | 6.21 | 128k | |
 Mi:dm K 2.5 Pro | -- | -- | -- | -- | 128k | |
| LFM2.5-2.6B | -- | -- | -- | -- | 128k | |
| LFM2.5-8B-A1B | -- | -- | -- | -- | 33k | |
| GPT-5.5 Pro Xhigh | -- | -- | -- | -- | 922k | |
| Nova Premier | -- | -- | -- | -- | 1M | |
 Cogito v2.1 Reasoning | -- | -- | -- | -- | 128k | |
| Ring-2.6-1T | -- | -- | -- | 12.82 | 262k | |
| Phi-4 | -- | -- | -- | 4.06 | 16k | |
| Agnes 2.5 Pro Beta | -- | -- | -- | -- | 1M | |
| Jamba Reasoning 3B | -- | -- | -- | -- | 262k | |
| Quasar 438B Max, Based on GLM-5.2 | -- | -- | -- | -- | 1M | |
| LFM2 24B A2B | -- | -- | -- | -- | 33k | |
 KAT Coder Pro V2 | -- | -- | -- | -- | 256k | |
| Qwen3 Coder Next | -- | -- | -- | 1.60 | 256k |
Key definitions
Frequently Asked Questions
Claude Fable 5.2 (Max, Default Fallback) currently ranks #1 on the Artificial Analysis LLM Leaderboard with an Intelligence Index score of 65, out of 113 models ranked.
The top models by Intelligence Index are: 1. Claude Fable 5.2 (Max, Default Fallback) (65), 2. Claude Fable 5.2 (Xhigh, Default Fallback) (62), 3. Claude Fable 5.2 (High, Default Fallback) (62), 4. Gemini 4 Carbon (High) (60), 5. Claude Opus 5.5 (Max, Default Fallback) (59).
GPT-6 Luna (Low) is the fastest, with the lowest time per task at 0.6 minutes per Intelligence Index task, followed by GPT-6.1 Terra (Low) (1.2 min) and Gemini 4 Hydrogen (Low) (1.4 min). Lower is better.
K2 Horizon 375B A23B has the lowest cost per Intelligence Index task at $0.00, followed by Command A+ ($0.00) and North Mini Code ($0.00).
Kimi K3.1 Pro is the highest-ranked open weights model with an Intelligence Index score of 46. There are 44 open weights models out of 113 total on the leaderboard.
The top open weights models by Intelligence Index are: 1. Kimi K3.1 Pro (46), 2. DeepSeek V4.1 Flash 1017 (Max) (33), 3. GLM-5.3 (Max) (32).
Claude Fable 5.2 (Max, Default Fallback) leads among 108 reasoning models with an Intelligence Index score of 65. Reasoning models use extended thinking to solve complex problems before responding.
The leaderboard includes filters to narrow results by model type (reasoning vs non-reasoning), openness (open weights vs proprietary), and other criteria.
Click on any model name in the leaderboard to visit its dedicated comparison page with detailed charts covering intelligence, pricing, speed, latency, and more. You can also compare API providers for each model. View all models