LLM Leaderboard & AI Model Rankings
One page that aggregates the major AI benchmarks and leaderboards — Chatbot Arena Elo (arena.ai), the Artificial Analysis Intelligence Index and speech arenas, and OpenRouter usage — across chat, coding, TTS, speech-to-text, image, video and agents.
| Rank | Model | Vendor | Elo ? | Votes ? | Input ? | Output | Cache ? |
|---|---|---|---|---|---|---|---|
| 1 |  gpt-image-2.5-sunburst | OpenAI | 1424 ±8 | 10,884 | — | — | — |
| 2 | gpt-image-2.5-flare | OpenAI | 1401 ±8 | 10,058 | — | — | — |
| 3 | gpt-image-2 (medium) | OpenAI | 1383 ±4 | 88,744 | — | — | — |
| 4 | Microsoft AI | 1335 ±6 | 18,625 | — | — | — | |
| 5 |  reve-2.1 | Reve | 1302 ±8 | 8,305 | — | — | — |
| 6 |  grok-imagine-image-2.0 (low) | SpaceXAI | 1301 ±8 | 7,257 | — | — | — |
| 7 |  muse-image | Meta | 1276 ±5 | 35,422 | — | — | — |
| 8 | reve-2.0 | Reve | 1269 ±6 | 15,802 | — | — | — |
| 9 |  gemini-3.1-flash-image (nano-banana-2) [web-search] | 1261 ±4 | 53,228 | 0.500 | 3.00 | — | |
| 10 |  seedream-5.0-pro | Bytedance | 1256 ±4 | 85,690 | — | — | — |
| 11 |  qwen-image-3.0-pro | Alibaba | 1256 ±6 | 11,520 | — | — | — |
| 12 | Microsoft AI | 1254 ±4 | 61,448 | — | — | — | |
| 13 | gemini-3.1-flash-lite-image (nano-banana-2-lite) | 1250 ±6 | 17,851 | 0.250 | 1.50 | — | |
| 14 | gemini-3-pro-image-2k (nano-banana-pro) | 1247 ±3 | 168,447 | — | — | — | |
| 15 | gpt-image-1.5-high-fidelity | OpenAI | 1238 ±3 | 165,437 | — | — | — |
| 16 | gemini-3-pro-image-preview (nano-banana-pro) | 1234 ±5 | 86,451 | 2.00 | 12.0 | 0.200 | |
| 17 | qwen-image-2.1 | Alibaba | 1228 ±9 | 4,445 | — | — | — |
| 18 |  ideogram-4.0-quality | Ideogram | 1205 ±4 | 53,371 | — | — | — |
| 19 | qwen-image-2.0-pro-2026-06-22 | Alibaba | 1192 ±6 | 13,123 | — | — | — |
| 20 |  uni-1.1-max | Luma AI | 1189 ±6 | 14,083 | — | — | — |
| 21 | Microsoft AI | 1184 ±5 | 51,477 | — | — | — | |
| 22 | uni-1.1 | Luma AI | 1183 ±4 | 41,412 | — | — | — |
| 23 | grok-imagine-image | SpaceXAI | 1170 ±3 | 262,132 | — | — | — |
| 24 |  recraft-v4.1-utility-pro | Recraft | 1169 ±11 | 2,618 | — | — | — |
| 25 |  Cosmos3-Super-Text2Image (Agentic) | Nvidia | 1165 ±7 | 10,856 | — | — | — |
| 26 | grok-imagine-image-pro | SpaceXAI | 1162 ±4 | 99,444 | — | — | — |
| 27 |  flux-2-max | Black Forest Labs | 1162 ±3 | 123,187 | — | — | — |
| 28 | flux-2-flex | Black Forest Labs | 1157 ±3 | 156,912 | — | — | — |
| 29 | reve-v1.5 | Reve | 1155 ±4 | 39,755 | — | — | — |
| 30 | flux-2-pro | Black Forest Labs | 1154 ±3 | 205,540 | — | — | — |
| 31 | Cosmos3-Super-Text2Image | Nvidia | 1152 ±6 | 18,581 | — | — | — |
| 32 | gemini-2.5-flash-image-preview (nano-banana) | 1150 ±2 | 878,451 | 0.300 | 2.50 | 0.030 | |
| 33 |  hunyuan-image-3.0 | Tencent | 1150 ±3 | 180,550 | — | — | — |
| 34 | imagen-ultra-4.0-generate-001 | 1148 ±4 | 388,772 | — | — | — | |
| 35 | seedream-4.5 | Bytedance | 1147 ±3 | 287,128 | — | — | — |
| 36 | flux-2-dev | Black Forest Labs | 1146 ±4 | 78,080 | — | — | — |
| 37 | seedream-4-2k | Bytedance | 1140 ±7 | 12,566 | — | — | — |
| 38 | seedream-5.0-lite | Bytedance | 1138 ±3 | 123,178 | — | — | — |
| 39 | wan2.6-t2i | Alibaba | 1137 ±3 | 223,535 | — | — | — |
| 40 | recraft-v4.1-pro | Recraft | 1133 ±10 | 2,795 | — | — | — |
| 41 | imagen-4.0-generate-001 | 1129 ±3 | 539,788 | — | — | — | |
| 42 | qwen-image-2512 | Alibaba | 1125 ±3 | 98,363 | — | — | — |
| 43 |  krea-2-medium | Krea | 1123 ±4 | 50,800 | — | — | — |
| 44 | seedream-4-fal | Bytedance | 1117 ±7 | 11,842 | — | — | — |
| 45 | wan2.5-t2i-preview | Alibaba | 1116 ±3 | 271,922 | — | — | — |
| 46 |  hidream-o1-image | HiDream | 1116 ±4 | 59,511 | — | — | — |
| 47 | gpt-image-1 | OpenAI | 1116 ±3 | 268,466 | — | — | — |
| 48 | recraft-v4 | Recraft | 1115 ±3 | 129,674 | — | — | — |
| 49 | seedream-4-high-res-fal | Bytedance | 1113 ±3 | 182,669 | — | — | — |
| 50 | krea-2-turbo | Krea | 1113 ±5 | 43,913 | — | — | — |
| 51 | krea-2-large | Krea | 1110 ±4 | 49,906 | — | — | — |
| 52 | gpt-image-1-mini | OpenAI | 1110 ±3 | 171,058 | — | — | — |
| 53 | wan2.7-image-pro | Alibaba | 1103 ±5 | 30,229 | — | — | — |
| 54 | wan2.7-image | Alibaba | 1101 ±5 | 30,621 | — | — | — |
| 55 | recraft-v4.1-flash | Recraft | 1100 ±9 | 3,533 | — | — | — |
| 56 | Microsoft AI | 1093 ±4 | 98,414 | — | — | — | |
| 57 | z-image-turbo | Alibaba | 1084 ±5 | 24,291 | — | — | — |
| 58 | seedream-3 | Bytedance | 1082 ±5 | 36,890 | — | — | — |
| 59 | flux-1-kontext-max | Black Forest Labs | 1074 ±3 | 65,517 | — | — | — |
| 60 | flux-2-klein-9b | Black Forest Labs | 1070 ±3 | 150,780 | — | — | — |
| 61 | qwen-image-prompt-extend | Alibaba | 1060 ±3 | 716,829 | — | — | — |
| 62 | flux-1-kontext-pro | Black Forest Labs | 1059 ±3 | 332,376 | — | — | — |
| 63 | imagen-3.0-generate-002 | 1058 ±3 | 359,786 | — | — | — | |
| 64 | qwen-image | Alibaba | 1057 ±3 | 84,555 | — | — | — |
| 65 | ideogram-v3-quality | Ideogram | 1048 ±4 | 118,542 | — | — | — |
| 66 | photon | Luma AI | 1036 ±4 | 130,158 | — | — | — |
| 67 | p-image | — | 1033 ±3 | 109,782 | — | — | — |
| 68 | flux-2-klein-4b | Black Forest Labs | 1030 ±3 | 152,358 | — | — | — |
| 69 |  runway-gen4 | Runway | 1024 ±4 | 54,376 | — | — | — |
| 70 | recraft-v3 | Recraft | 1021 ±4 | 195,623 | — | — | — |
| 71 | flux-1.1-pro | Black Forest Labs | 1016 ±3 | 70,460 | — | — | — |
| 72 | ideogram-v2 | Ideogram | 1013 ±4 | 72,090 | — | — | — |
| 73 |  lucid-origin | Leonardo AI | 1013 ±3 | 287,345 | — | — | — |
| 74 |  glm-image | Z.ai | 1012 ±9 | 4,852 | — | — | — |
| 75 | gemini-2.0-flash-preview-image-generation | 975 ±3 | 257,241 | — | — | — | |
| 76 | flux-1-dev-fp8 | Black Forest Labs | 969 ±4 | 49,221 | — | — | — |
| 77 | dall-e-3 | OpenAI | 968 ±4 | 238,749 | — | — | — |
| 78 | flux-1-kontext-dev | Black Forest Labs | 941 ±4 | 216,238 | — | — | — |
| 79 | stable-diffusion-v35-large | — | 938 ±5 | 23,396 | — | — | — |
| 80 | bagel | Bytedance | 898 ±6 | 12,363 | — | — | — |
Source: arena.ai · Updated Sep 24, 2026 · PandaNpc Updated Sep 28, 2026, 06:00 UTCView full leaderboard →
How to read these numbers
An LLM leaderboard tells you which model people prefer; an AI benchmark tells you what a model can objectively do; usage data tells you what developers actually ship. No single ranking answers "which is the best AI model", so this page aggregates all three — Chatbot Arena Elo from arena.ai, the Artificial Analysis Intelligence Index and speech arenas, and OpenRouter token share — and refreshes them daily.
arena.ai (formerly LMArena) shows users two anonymous answers to the same prompt and asks them to pick one, then derives an Elo rating from those battles. It measures which answer people prefer — not objective capability. A likeable style can win, while a rigorous but verbose answer can lose.
Votes are the battles a model has accumulated. A newly listed model has few votes and a wide confidence interval, so its rank swings easily — always read the rank together with the ± interval.
The Agent board works differently from the rest: it measures real agent sessions rather than blind-vote preference. Its headline metric, Net Improvement, is a percentage gain — not an Elo rating — so it cannot be compared across boards. arena.ai also reports five more dimensions (Confirmed Success, Steerability, Bash Recovery and others) on the full leaderboard.
The usage board comes from OpenRouter and counts real API token share — developers voting with their wallets, which complements the subjective preference boards.
The speech boards come from Artificial Analysis: TTS uses blind-listening Speech Arena Elo; ASR uses AA-WER v2, where lower is better; and native Speech-to-Speech uses an equal-weight composite of Big Bench Audio, Full Duplex Bench and τ-Voice, where higher is better. Compare scores only within the same board. Source: artificialanalysis.ai.
The Intelligence Index board comes from Artificial Analysis: it distills many public benchmarks into a single 0-100 Intelligence Index, measuring objective capability rather than subjective preference — a natural complement to arena.ai's human blind vote. Source: artificialanalysis.ai.