Head-to-Head Leaderboard
A persistent cross-model leaderboard ranking AI models by Elo and win rate across every head-to-head matchup, with separate ladders for text, image, and video.
Text models
| Rank | Model | Elo | Win rate | Matches |
|---|---|---|---|---|
| 1 | grok-4.3 | 1706 | 96% | 23 |
| 2 | Kimi-K2.7-Code | 1570 | 81% | 8 |
| 3 | Meta: Muse Spark 1.1 | 1558 | 90% | 5 |
| 4 | Google: Gemini 3.7 Flash | 1532 | 100% | 2 |
| 5 | grok-4.6 | 1531 | 100% | 2 |
| 6 | gpt-oss-120b | 1529 | 64% | 7 |
| 7 | DeepSeek-V3.2 | 1528 | 100% | 2 |
| 8 | Google: Gemini 3.6 Flash | 1524 | 75% | 4 |
| 9 | Google: Gemini 3.5 Flash | 1517 | 100% | 1 |
| 10 | Anthropic: Claude Sonnet 4.6 | 1516 | 100% | 1 |
| 11 | DeepSeek-R1 | 1516 | 100% | 1 |
| 12 | nemotron-3-ultra-550b-a55b | 1516 | 100% | 1 |
| 13 | DeepSeek: DeepSeek V4 Pro | 1516 | 100% | 1 |
| 14 | OpenAI: GPT-5.6 Sol Pro | 1516 | 100% | 1 |
| 15 | MoonshotAI: Kimi K3 | 1516 | 100% | 1 |
| 16 | diffusiongemma-26b-a4b-it | 1516 | 100% | 1 |
| 17 | OpenAI: GPT-5.6 Luna | 1516 | 75% | 2 |
| 18 | DeepSeek-V4-Flash-0731 | 1516 | 75% | 2 |
| 19 | OpenRouter: Fusion | 1515 | 100% | 1 |
| 20 | OpenAI: GPT-5.6 Sol | 1515 | 100% | 1 |
| 21 | Kimi K3 | 1515 | 100% | 1 |
| 22 | OpenAI: GPT-5 Pro | 1515 | 100% | 1 |
| 23 | SpaceXAI: Grok 4.6 | 1515 | 100% | 1 |
| 24 | Mistral-Large-3 | 1515 | 100% | 1 |
| 25 | Sakana: Fugu Ultra | 1514 | 100% | 1 |
| 26 | DeepSeek: DeepSeek V4 Flash | 1501 | 50% | 2 |
| 27 | Anthropic: Claude Fable 5 | 1500 | 50% | 1 |
| 28 | Cohere-command-a-plus-05-2026 | 1499 | 50% | 2 |
| 29 | Kimi-K2.6 | 1493 | 33% | 3 |
| 30 | mistral-small-2503 | 1492 | 0% | 1 |
| 31 | mistral-medium-2505 | 1492 | 0% | 1 |
| 32 | gpt-5.4 | 1492 | 25% | 2 |
| 33 | Llama-3.3-70B-Instruct | 1492 | 0% | 1 |
| 34 | Kimi K3 | 1490 | 38% | 4 |
| 35 | DeepSeek-V4-Flash | 1490 | 40% | 5 |
| 36 | DeepSeek-V4-Pro | 1487 | 41% | 11 |
| 37 | Qwen: Qwen3.7 Flash | 1485 | 0% | 1 |
| 38 | gemma-4-31b-it | 1484 | 0% | 1 |
| 39 | OpenAI: GPT-5.5 Pro | 1484 | 0% | 1 |
| 40 | Kimi K2.7 Code | 1484 | 0% | 1 |
| 41 | xAI: Grok 4.5 | 1484 | 0% | 1 |
| 42 | Grok 4.20 0309 Non Reasoning | 1484 | 0% | 1 |
| 43 | Xiaomi: MiMo-V2.5 | 1483 | 0% | 1 |
| 44 | Sarvam M | 1482 | 0% | 1 |
| 45 | mistral-medium-3-5 | 1479 | 0% | 2 |
| 46 | Codestral-2501 | 1478 | 25% | 4 |
| 47 | cohere-command-a | 1476 | 25% | 4 |
| 48 | Ministral-3B | 1476 | 0% | 2 |
| 49 | gpt-5.4-mini | 1475 | 0% | 2 |
| 50 | Phi-4-multimodal-instruct | 1475 | 0% | 2 |
| 51 | gpt-5.4-nano | 1474 | 0% | 2 |
| 52 | Llama-4-Maverick-17B-128E-Instruct-FP8 | 1472 | 0% | 2 |
| 53 | Z.ai: GLM 5.3 | 1469 | 0% | 2 |
| 54 | gpt-oss-20b | 1469 | 0% | 2 |
| 55 | Anthropic: Claude Opus 4.8 | 1468 | 25% | 4 |
| 56 | Phi-4-reasoning | 1458 | 0% | 3 |
| 57 | Llama-4-Scout-17B-16E-Instruct | 1457 | 0% | 3 |
| 58 | Phi-4 | 1448 | 0% | 4 |
| 59 | Phi-4-mini-instruct | 1444 | 0% | 4 |
| 60 | GLM 5.2 | 1442 | 10% | 5 |
Image models
| Rank | Model | Elo | Win rate | Matches |
|---|---|---|---|---|
| 1 | GPT Image 2 API | 1587 | 100% | 7 |
| 2 | Fibo Bbq Preview | 1563 | 100% | 5 |
| 3 | Luma Uni-1 Edit Max | 1559 | 100% | 5 |
| 4 | Imagineart 2.0 Preview | 1555 | 100% | 5 |
| 5 | Fibo | 1544 | 90% | 5 |
| 6 | Rundiffusion Photo Flux | 1544 | 100% | 4 |
| 7 | Ideogram V4.0 Text to Image | 1542 | 90% | 5 |
| 8 | ImagineArt 1.5 Pro Preview | 1540 | 83% | 6 |
| 9 | Fibo Lite | 1534 | 75% | 6 |
| 10 | Qwen Image 3 Image Editing | 1531 | 100% | 2 |
| 11 | Grok Imagine Image 2.0 | 1531 | 100% | 2 |
| 12 | Fibo Gen 1.5 Text to Image | 1531 | 100% | 2 |
| 13 | Juggernaut Flux Lightning | 1531 | 80% | 5 |
| 14 | Krea 2 Large | 1530 | 100% | 3 |
| 15 | Reve 2.1 | 1523 | 100% | 2 |
| 16 | Nano Banana Lite Edit | 1520 | 100% | 2 |
| 17 | FLUX 2 | 1518 | 100% | 1 |
| 18 | FLUX.1 Krea [dev] | 1516 | 100% | 1 |
| 19 | V4.0q [instant] | 1508 | 100% | 1 |
| 20 | Seedream 5.0 Pro Image Editing | 1508 | 60% | 5 |
| 21 | Cosmos 3 Super | 1503 | 70% | 5 |
| 22 | FLUX.1 Krea [dev] with LoRAs | 1502 | 50% | 1 |
| 23 | Bitdance | 1501 | 50% | 1 |
| 24 | Bria RMBG 2.0 | 1500 | 50% | 1 |
| 25 | Grok Imagine Image Quality | 1487 | 0% | 1 |
| 26 | FLUX.1 [schnell] | 1486 | 0% | 1 |
| 27 | FLUX.1 Krea [dev] | 1486 | 0% | 1 |
| 28 | FLUX.1 Kontext [pro] | 1486 | 0% | 1 |
| 29 | DeepSeek Janus-Pro | 1486 | 0% | 1 |
| 30 | Ernie Image Lora | 1485 | 0% | 1 |
| 31 | Bytedance Seedream V4.5 Edit | 1485 | 0% | 1 |
| 32 | FLUX.1 [dev] with LoRAs | 1485 | 0% | 1 |
| 33 | Glm Image | 1485 | 0% | 1 |
| 34 | FLUX.2 [klein] 9B | 1485 | 0% | 1 |
| 35 | Ernie Image Turbo | 1485 | 0% | 1 |
| 36 | FLUX 2 Lora Gallery Realism | 1485 | 0% | 1 |
| 37 | FLUX1.1 [pro] | 1485 | 0% | 1 |
| 38 | FLUX.1 SRPO [dev] | 1485 | 0% | 1 |
| 39 | FLUX.1 [dev] with Controlnets and Loras | 1484 | 0% | 1 |
| 40 | Boogu Image | 1484 | 0% | 1 |
| 41 | Ernie Image Lora Turbo | 1484 | 0% | 1 |
| 42 | FLUX.2 [klein] 9B Base LoRA | 1484 | 0% | 1 |
| 43 | CogView | 1484 | 0% | 1 |
| 44 | FLUX.1 SRPO [dev] | 1484 | 0% | 1 |
| 45 | Juggernaut Flux Base LoRA | 1477 | 50% | 5 |
| 46 | Bagel | 1318 | 9% | 23 |
| 47 | AuraFlow | 1284 | 10% | 30 |
Video models
| Rank | Model | Elo | Win rate | Matches |
|---|---|---|---|---|
| 1 | Happy Horse | 1623 | 100% | 10 |
| 2 | Seedance 2 Image to Video | 1591 | 94% | 9 |
| 3 | Happy Horse 1.1 Image to Video | 1588 | 100% | 7 |
| 4 | Gemini Omni Flash | 1561 | 100% | 5 |
| 5 | Luma Ray 3.2 Image to Video | 1534 | 72% | 9 |
| 6 | Flux 3 Image To Video Draft | 1531 | 100% | 2 |
| 7 | MiniMax H3 Reference to Video | 1531 | 100% | 2 |
| 8 | Seedance 2.5 Reference to Video | 1531 | 100% | 2 |
| 9 | LTX 2.5 Image to Video Pro | 1516 | 75% | 2 |
| 10 | Flux 3 Image to Video | 1513 | 75% | 2 |
| 11 | Wan v2.6 Image to Video | 1511 | 64% | 7 |
| 12 | Marey Realism V1.5 | 1510 | 61% | 9 |
| 13 | Wan 3.0 | 1510 | 100% | 1 |
| 14 | Bytedance Seedance V1.5 Pro Image To Video | 1510 | 50% | 7 |
| 15 | Kling Video v3 Image to Video [Pro] | 1502 | 50% | 1 |
| 16 | Kling Video V3 Turbo Pro Image to Video | 1501 | 50% | 1 |
| 17 | Krea Wan 14b- Text to Video | 1500 | 50% | 1 |
| 18 | LTX 2.5 Image to Video Pro | 1500 | 50% | 2 |
| 19 | Kandinsky5 | 1500 | 50% | 1 |
| 20 | Kling O3 Image to Video [Pro] | 1500 | 50% | 1 |
| 21 | Kling Video | 1500 | 50% | 1 |
| 22 | Kling Video v2.6 Image to Video | 1489 | 0% | 1 |
| 23 | Bernini-R Edit Video | 1489 | 38% | 12 |
| 24 | LTX Video (preview) | 1488 | 0% | 1 |
| 25 | Kling 1.6 | 1487 | 0% | 1 |
| 26 | Heygen | 1485 | 0% | 1 |
| 27 | Infinity Star | 1485 | 0% | 1 |
| 28 | Kandinsky5 | 1485 | 0% | 1 |
| 29 | Kling 2.0 Master | 1485 | 0% | 1 |
| 30 | Veo 2 (Image to Video) | 1484 | 0% | 1 |
| 31 | Heygen Video Agent | 1484 | 0% | 1 |
| 32 | Heygen v5 Digital Twin | 1484 | 0% | 1 |
| 33 | Cosmos Predict 2.5 2B | 1484 | 0% | 1 |
| 34 | Kandinsky5 Pro | 1484 | 0% | 1 |
| 35 | CogVideoX-5B | 1421 | 0% | 6 |
| 36 | AnimateDiff Turbo | 1352 | 0% | 13 |
| 37 | AnimateDiff | 1350 | 0% | 13 |
Anim models
| Rank | Model | Elo | Win rate | Matches |
|---|---|---|---|---|
| 1 | Kimi K3 | 1531 | 100% | 2 |
| 2 | mistral-medium-3-5 | 1516 | 100% | 1 |
| 3 | Anthropic: Claude Opus 4.8 | 1485 | 0% | 1 |
| 4 | Kimi-K2.7-Code | 1484 | 0% | 1 |
| 5 | xAI: Grok 4.5 | 1484 | 0% | 1 |
See every matchup on the Head-to-Head index.