LLM Model Rankings Live
Live rankings of 1,069 large language models from 178 providers, aggregated from LMSys Chatbot Arena, OpenRouter, Artificial Analysis, and the Hugging Face open-source leaderboard. Updated every 6 hours — last refresh 35d ago.
Powered by:
LMSys Arena
OpenRouter
Artificial Analysis
Hugging Face
Models tracked
1,069
20 with usage data
Providers
178
All major labs
#1 Arena ELO
claude-opus-4-6-search
1251 ELO
#1 Token share
Hy3 preview
8.9% weekly tokens
OpenRouter — Top 15 by Weekly Token Share
Real production usage on OpenRouter — weekly token volume across all inference providers. Matches OpenRouter's official rankings page.
LMSys Arena — Top 15 by ELO
Human-voted preference ratings from blind side-by-side comparisons on Chatbot Arena.
Artificial Analysis — Quality Index
Independent benchmark suite covering reasoning, knowledge, code, and instruction-following.
Token Share by Provider
Aggregated OpenRouter weekly tokens grouped by provider — who owns inference traffic right now.
Quality vs. Price
Bubble chart — find the best value: top-left is high quality at low price (log scale).
Quality vs. Speed (Tokens/sec)
Speed-quality trade-off — right side is faster, top side is smarter.
Input vs. Output Pricing
All priced models — log/log scatter. Diagonal line = equal input/output price.
Context Length Distribution
How models cluster by maximum context window — most production traffic still lives at ≤128K.
Open-source vs. Proprietary
Token-weighted market share — what % of OpenRouter traffic flows to open-weights models.
Full leaderboard
All 1,069 ranked models, sortable by any column. Click a row to see history and full specs.
| Model | Provider | Arena # | ELO | OR # | Quality | Context | $/1M in | $/1M out |
|---|---|---|---|---|---|---|---|---|
| claude-opus-4-6-search | — | 1 | 1251 | — | — | — | — | — |
| claude-opus-4-6-thinking | — | 1 | 1502 | — | — | — | — | — |
| dreamina-seedance-2.0-720p | — | 1 | 1457 | — | — | — | — | — |
| gpt-image-2 (medium) | — | 1 | 1389 | — | — | — | — | — |
| Hy3 preview OS | tencent | — | — | 1 | 42 | 262K | $0.063 | $0.21 |
| chatgpt-image-latest-high-fidelity (20251216) | — | 2 | 1392 | — | — | — | — | — |
| claude-opus-4-7-thinking | — | 2 | 1500 | — | — | — | — | — |
| DeepSeek V4 Flash | deepseek | — | — | 2 | 36 | 1.0M | $0.098 | $0.20 |
| gemini-3.1-flash-image-preview (nano-banana-2) [web-search] | — | 2 | 1270 | — | — | — | — | — |
| gpt-5.5-search | — | 2 | 1239 | — | — | — | — | — |
| happyhorse-1.0 | — | 2 | 1435 | — | — | — | — | — |
| claude-opus-4-6 | — | 3 | 1498 | — | — | — | — | — |
| gemini-3-pro-image-preview-2k (nano-banana-pro) | — | 3 | 1243 | — | — | — | — | — |
| grok-imagine-image-quality (20260519) | — | 3 | 1388 | — | — | — | — | — |
| grok-imagine-video | — | 3 | 1265 | — | — | — | — | — |
| MiMo-V2.5 MM | xiaomi | — | — | 3 | 49 | 1.0M | $0.14 | $0.28 |
| veo-3.1-audio-1080p | — | 3 | 1372 | — | — | — | — | — |
| claude-opus-4-7 | — | 4 | 1492 | — | — | — | — | — |
| ernie-5.1 | — | 4 | 1226 | — | — | — | — | — |
| gpt-image-1.5-high-fidelity | — | 4 | 1241 | — | — | — | — | — |
| kling-o3-pro | — | 4 | 1247 | — | — | — | — | — |
| Owl Alpha | openrouter | — | — | 4 | — | 1.0M | — | — |
| sora-2-pro | — | 4 | 1368 | — | — | — | — | — |
| Claude Sonnet 4.6 MM | anthropic | — | — | 5 | — | 1.0M | $3.00 | $15.0 |
| claude-sonnet-4-6-search | — | 5 | 1219 | — | — | — | — | — |
| gemini-3-pro-image-preview (nano-banana-pro) | — | 5 | 1232 | — | — | — | — | — |
| glm-5.1 | — | 5 | 1532 | — | — | — | — | — |
| kling-o1-pro | — | 5 | 1208 | — | — | — | — | — |
| muse-spark | — | 5 | 1489 | — | — | — | — | — |
| veo-3.1-audio | — | 5 | 1366 | — | — | — | — | — |
| Claude Opus 4.7 MM | anthropic | — | — | 6 | — | 1.0M | $5.00 | $25.0 |
| claude-sonnet-4-6 | — | 6 | 1524 | — | — | — | — | — |
| gemini-3.1-pro-grounding | — | 6 | 1216 | — | — | — | — | — |
| gemini-3.1-pro-preview | — | 6 | 1488 | — | — | — | — | — |
| grok-imagine-image-quality | — | 6 | 1228 | — | — | — | — | — |
| runway-gen4-aleph | — | 6 | 1204 | — | — | — | — | — |
| veo-3.1-fast-audio | — | 6 | 1364 | — | — | — | — | — |
| DeepSeek V4 Pro | deepseek | — | — | 7 | 39 | 1.0M | $0.43 | $0.87 |
| gemini-3-pro | — | 7 | 1486 | — | — | — | — | — |
| gpt-5.2-search | — | 7 | 1210 | — | — | — | — | — |
| gpt-5.5 | — | 7 | 1288 | — | — | — | — | — |
| grok-imagine-video-480p | — | 7 | 1383 | — | — | — | — | — |
| kimi-k2.6 | — | 7 | 1519 | — | — | — | — | — |
| uni-1.1-max | — | 7 | 1192 | — | — | — | — | — |
| veo-3.1-fast-audio-1080p | — | 7 | 1364 | — | — | — | — | — |
| gpt-5.2-chat-latest-20260210 | — | 8 | 1280 | — | — | — | — | — |
| gpt-5.4 | — | 8 | 1474 | — | — | — | — | — |
| gpt-5.5-high | — | 8 | 1481 | — | — | — | — | — |
| gpt-5.5-high (codex-harness) | — | 8 | 1519 | — | — | — | — | — |
| grok-4.20-multi-agent-beta-0309 | — | 8 | 1209 | — | — | — | — | — |
| grok-imagine-video-720p | — | 8 | 1357 | — | — | — | — | — |
| mai-image-2 | — | 8 | 1182 | — | — | — | — | — |
| MiMo-V2.5-Pro | xiaomi | — | — | 8 | 54 | 1.0M | $0.43 | $0.87 |
| claude-opus-4-5-20251101 | — | 9 | 1466 | — | — | — | — | — |
| DeepSeek V3.2 | deepseek | — | — | 9 | — | 131K | $0.23 | $0.34 |
| gemini-3-pro-grounding | — | 9 | 1208 | — | — | — | — | — |
| gemini-3.5-flash | — | 9 | 1480 | — | — | — | — | — |
| uni-1.1 | — | 9 | 1176 | — | — | — | — | — |
| veo-3-fast-audio | — | 9 | 1349 | — | — | — | — | — |
| vidu-q3-pro | — | 9 | 1361 | — | — | — | — | — |
| Gemini 3 Flash Preview MM | — | — | 10 | — | 1.0M | $0.50 | $3.00 | |
| gemini-3-flash-grounding | — | 10 | 1206 | — | — | — | — | — |
| gpt-5.4-high | — | 10 | 1480 | — | — | — | — | — |
| gpt-5.5 (codex-harness) | — | 10 | 1489 | — | — | — | — | — |
| gpt-5.5-xhigh (codex-harness) | — | 10 | 1503 | — | — | — | — | — |
| grok-imagine-image | — | 10 | 1173 | — | — | — | — | — |
| kling-v3-pro | — | 10 | 1360 | — | — | — | — | — |
| wan2.6-t2v | — | 10 | 1341 | — | — | — | — | — |
| Nemotron 3 Super | nvidia | — | — | 11 | — | 1.0M | $0.090 | $0.45 |
| Gemini 2.5 Flash Lite MM | — | — | 12 | — | 1.0M | $0.100 | $0.40 | |
| Claude Opus 4.8 MM | anthropic | — | — | 13 | — | 1.0M | $5.00 | $25.0 |
| gpt-oss-120b OS | openai | — | — | 13 | — | 131K | $0.039 | $0.18 |
| Gemini 2.5 Flash MM | — | — | 14 | — | 1.0M | $0.30 | $2.50 | |
| Laguna M.1 (free) | poolside | — | — | 16 | — | 262K | — | — |
| Gemini 3.5 Flash MM | — | — | 17 | 55 | 1.0M | $1.50 | $9.00 | |
| Claude Opus 4.6 MM | anthropic | — | — | 18 | — | 1.0M | $5.00 | $25.0 |
| GPT-4o-mini MM | openai | — | — | 19 | — | 128K | $0.15 | $0.60 |
| Kimi K2.6 MM | moonshotai | — | — | 19 | — | 262K | $0.68 | $3.42 |
| MiniMax M2.7 | minimax | — | — | 21 | 50 | 205K | $0.28 | $1.20 |
| GPT-5.5 MM | openai | — | — | 22 | — | 1.1M | $5.00 | $30.0 |
| Gemini 3.1 Flash Lite MM | — | — | 23 | — | 1.0M | $0.25 | $1.50 | |
| Gemma 4 31B MM | — | — | 24 | — | 262K | $0.12 | $0.37 | |
| MiniMax M3 MM | minimax | — | — | 25 | — | 1.0M | $0.30 | $1.20 |
| Step 3.7 Flash MM | stepfun | — | — | 26 | — | 256K | $0.20 | $1.15 |
| GLM 5.1 | z ai | — | — | 27 | 51 | 203K | $0.98 | $3.08 |
| Claude Haiku 4.5 MM | anthropic | — | — | 28 | — | 200K | $1.00 | $5.00 |
| Gemini 3.1 Pro Preview MM | — | — | 28 | 57 | 1.0M | $2.00 | $12.0 | |
| GPT-5.4 MM | openai | — | — | 29 | — | 1.1M | $2.50 | $15.0 |
| Gemma 4 26B A4B MM | — | — | 31 | — | 262K | $0.060 | $0.33 | |
| GLM 4.5 Air | z ai | — | — | 31 | — | 131K | $0.13 | $0.85 |
| GPT-5 Mini MM | openai | — | — | 33 | — | 400K | $0.25 | $2.00 |
| Gemini 3.1 Flash Lite Preview MM | — | — | 34 | 34 | 1.0M | $0.25 | $1.50 | |
| Qwen3.7 Max | qwen | — | — | 35 | — | 1.0M | $1.25 | $3.75 |
| MiniMax M2.5 | minimax | — | — | 36 | — | 205K | $0.15 | $1.15 |
| Mistral Nemo | mistralai | — | — | 37 | — | 131K | $0.020 | $0.030 |
| Qwen3.6 Plus MM | qwen | — | — | 37 | — | 1.0M | $0.33 | $1.95 |
| Claude Sonnet 4.5 MM | anthropic | — | — | 38 | — | 1.0M | $3.00 | $15.0 |
| GLM 4.7 | z ai | — | — | 39 | — | 203K | $0.40 | $1.75 |
| Qwen3 235B A22B Instruct 2507 | qwen | — | — | 39 | — | 262K | $0.071 | $0.100 |
| GPT-5.4 Mini MM | openai | — | — | 40 | 23 | 400K | $0.75 | $4.50 |
| gpt-oss-20b OS | openai | — | — | 42 | — | 131K | $0.030 | $0.14 |
| Kimi K2.5 MM | moonshotai | — | — | 42 | — | 262K | $0.40 | $1.90 |
| GLM 5 | z ai | — | — | 44 | 50 | 203K | $0.60 | $1.92 |
| Llama 3.1 8B Instruct OS | meta llama | — | — | 46 | — | 131K | $0.020 | $0.050 |
| Qwen3.5-Flash MM | qwen | — | — | 46 | — | 1.0M | $0.065 | $0.26 |
| GPT-5 MM | openai | — | — | 47 | — | 400K | $1.25 | $10.0 |
| GPT-4.1 Mini MM | openai | — | — | 48 | — | 1.0M | $0.40 | $1.60 |
| GPT-5.3-Codex MM | openai | — | — | 48 | — | 400K | $1.75 | $14.0 |
| Laguna XS.2 (free) OS | poolside | — | — | 50 | — | 262K | — | — |
| Gemini 2.0 Flash MM | — | — | 51 | — | 1.0M | $0.100 | $0.40 | |
| GPT-5.4 Nano MM | openai | — | — | 51 | 38 | 400K | $0.20 | $1.25 |
| Gemini 2.5 Pro MM | — | — | 52 | 35 | 1.0M | $1.25 | $10.0 | |
| Grok 4.3 MM | x ai | — | — | 52 | — | 1.0M | $1.25 | $2.50 |
| DeepSeek V3 0324 | deepseek | — | — | 54 | — | 164K | $0.20 | $0.77 |
| GPT-5 Nano MM | openai | — | — | 55 | — | 400K | $0.050 | $0.40 |
| Ling-2.6-flash | inclusionai | — | — | 56 | 26 | 262K | $0.010 | $0.030 |
| Qwen3 Embedding 8B MM | qwen | — | — | 56 | — | 32K | $0.010 | — |
| Qwen3.5-35B-A3B MM | qwen | — | — | 58 | — | 262K | $0.14 | $1.00 |
| Text Embedding 3 Small MM | openai | — | — | 58 | — | 8K | $0.020 | — |
| DeepSeek V3.1 | deepseek | — | — | 59 | — | 164K | $0.21 | $0.79 |
| MiMo-V2-Flash | xiaomi | — | — | 59 | 30 | 262K | $0.100 | $0.30 |
| Claude Opus 4.5 MM | anthropic | — | — | 63 | — | 200K | $5.00 | $25.0 |
| GPT-5.1 MM | openai | — | — | 64 | — | 400K | $1.25 | $10.0 |
| Nemotron 3 Nano 30B A3B | nvidia | — | — | 64 | — | 262K | $0.050 | $0.20 |
| GPT-4.1 MM | openai | — | — | 68 | — | 1.0M | $2.00 | $8.00 |
| Claude Sonnet 4 MM | anthropic | — | — | 70 | — | 1.0M | $3.00 | $15.0 |
| GPT-5.2 MM | openai | — | — | 70 | — | 400K | $1.75 | $14.0 |
| Qwen3.5 397B A17B MM | qwen | — | — | 71 | — | 262K | $0.39 | $2.34 |
| GLM 4.7 Flash | z ai | — | — | 72 | — | 203K | $0.060 | $0.40 |
| Gemini 2.0 Flash Lite MM | — | — | 73 | — | 1.0M | $0.075 | $0.30 | |
| Qwen3.6 35B A3B MM | qwen | — | — | 73 | — | 262K | $0.14 | $1.00 |
| GPT-4.1 Nano MM | openai | — | — | 74 | — | 1.0M | $0.100 | $0.40 |
| Qwen3 Coder 480B A35B | qwen | — | — | 75 | — | 1.0M | $0.22 | $1.80 |
| Qwen3.6 27B MM | qwen | — | — | 75 | — | 262K | $0.29 | $3.20 |
| Llama 4 Maverick MM | meta llama | — | — | 76 | — | 1.0M | $0.15 | $0.60 |
| Mistral Small 4 MM | mistralai | — | — | 77 | — | 262K | $0.15 | $0.60 |
| Mistral Small 3.2 24B MM | mistralai | — | — | 78 | — | 128K | $0.075 | $0.20 |
| Qwen3.5 Plus 2026-02-15 MM | qwen | — | — | 79 | — | 1.0M | $0.26 | $1.56 |
| Gemini 2.5 Flash Lite Preview 09-2025 MM | — | — | 80 | — | 1.0M | $0.100 | $0.40 | |
| DeepSeek V3.1 Terminus | deepseek | — | — | 81 | — | 164K | $0.27 | $0.95 |
| Qwen3.5-9B MM | qwen | — | — | 82 | — | 262K | $0.040 | $0.15 |
| Llama 3.3 70B Instruct OS | meta llama | — | — | 83 | — | 131K | $0.100 | $0.32 |
| Ministral 3 8B 2512 MM | mistralai | — | — | 83 | — | 262K | $0.15 | $0.15 |
| CoBuddy (free) | baidu | — | — | 84 | — | 131K | — | — |
| Qwen3.6 Flash MM | qwen | — | — | 84 | — | 1.0M | $0.19 | $1.13 |
| Claude Opus 4.8 (Fast) MM | anthropic | — | — | 85 | — | 1.0M | $10.0 | $50.0 |
| Qwen3 Next 80B A3B Instruct | qwen | — | — | 85 | — | 262K | $0.090 | $1.10 |
| GLM 5 Turbo | z ai | — | — | 86 | 47 | 203K | $1.20 | $4.00 |
| MiMo-V2-Pro | xiaomi | — | — | 88 | — | 1.0M | $1.00 | $3.00 |
| Ring-2.6-1T OS | inclusionai | — | — | 89 | 38 | 262K | $0.075 | $0.63 |
| Gemini Embedding 001 MM | — | — | 90 | — | 20K | $0.15 | — | |
| Qwen3.5-27B MM | qwen | — | — | 91 | — | 262K | $0.20 | $1.56 |
| Qwen3 32B | qwen | — | — | 92 | — | 131K | $0.080 | $0.28 |
| Step 3.5 Flash | stepfun | — | — | 92 | — | 262K | $0.090 | $0.30 |
| Gemma 3 27B MM | — | — | 93 | — | 131K | $0.080 | $0.16 | |
| Nemotron 3 Nano Omni (free) MM | nvidia | — | — | 94 | 21 | 256K | — | — |
| Qwen3 VL 32B Instruct MM | qwen | — | — | 94 | — | 262K | $0.10 | $0.42 |
| Qwen3 Coder Next OS | qwen | — | — | 95 | — | 262K | $0.11 | $0.80 |
| Kimi K2 0905 | moonshotai | — | — | 96 | — | 262K | $0.60 | $2.50 |
| Grok 4.20 MM | x ai | — | — | 97 | — | 2.0M | $1.25 | $2.50 |
| GPT-5.1 Chat MM | openai | — | — | 99 | — | 128K | $1.25 | $10.0 |
| DeepSeek V3.2 Exp | deepseek | — | — | 101 | — | 164K | $0.27 | $0.41 |
| Grok Build 0.1 MM | x ai | — | — | 102 | — | 256K | $1.00 | $2.00 |
| DeepSeek V3 | deepseek | — | — | 103 | — | 164K | $0.32 | $0.89 |
| Qwen3 Embedding 4B MM | qwen | — | — | 104 | — | 33K | $0.020 | — |
| Qwen3 30B A3B Instruct 2507 | qwen | — | — | 105 | — | 262K | $0.090 | $0.30 |
| Qwen3 VL 235B A22B Instruct MM | qwen | — | — | 109 | — | 262K | $0.20 | $0.88 |
| GPT-4o MM | openai | — | — | 110 | — | 128K | $2.50 | $10.0 |
| Mistral Large 3 2512 MM | mistralai | — | — | 110 | — | 262K | $0.50 | $1.50 |
| Claude 3.5 Haiku MM | anthropic | — | — | 112 | — | 200K | $0.80 | $4.00 |
| Llama 4 Scout MM | meta llama | — | — | 113 | — | 10.0M | $0.080 | $0.30 |
| Qwen3 8B OS | qwen | — | — | 114 | — | 131K | $0.050 | $0.40 |
| R1 0528 | deepseek | — | — | 115 | 27 | 164K | $0.50 | $2.15 |
| Nemotron Nano 9B V2 | nvidia | — | — | 116 | — | 131K | $0.040 | $0.16 |
| Text Embedding 3 Large MM | openai | — | — | 117 | — | 8K | $0.13 | — |
| Llama 3.1 70B Instruct | meta llama | — | — | 118 | — | 131K | $0.40 | $0.40 |
| Gemma 3 12B MM | — | — | 119 | — | 131K | $0.040 | $0.13 | |
| MiMo-V2-Omni MM | xiaomi | — | — | 119 | 43 | 262K | $0.40 | $2.00 |
| Qwen3 VL 8B Instruct MM | qwen | — | — | 120 | — | 256K | $0.080 | $0.50 |
| GLM 4.6 | z ai | — | — | 121 | — | 203K | $0.43 | $1.74 |
| Mistral Medium 3.5 MM | mistralai | — | — | 121 | — | 262K | $1.50 | $7.50 |
| Nano Banana (Gemini 2.5 Flash Image) MM | — | — | 122 | — | 33K | $0.30 | $2.50 | |
| o3 Mini MM | openai | — | — | 123 | — | 200K | $1.10 | $4.40 |
| GPT-5.3 Chat MM | openai | — | — | 124 | — | 128K | $1.75 | $14.0 |
| Qwen3 VL 30B A3B Instruct MM | qwen | — | — | 124 | — | 262K | $0.13 | $0.52 |
| bge-m3 MM | baai | — | — | 125 | — | 8K | $0.010 | — |
| Devstral 2 2512 MM | mistralai | — | — | 125 | — | 262K | $0.40 | $2.00 |
| Gemini 3.1 Pro Preview Custom Tools MM | — | — | 126 | — | 1.0M | $2.00 | $12.0 | |
| Llama 3 8B Lunaris | sao10k | — | — | 127 | — | 8K | $0.040 | $0.050 |
| GLM 5V Turbo MM | z ai | — | — | 128 | 43 | 203K | $1.20 | $4.00 |
| Nemotron Nano 12B 2 VL (free) MM | nvidia | — | — | 128 | — | 128K | — | — |
| Embed V1 4B MM | perplexity | — | — | 129 | — | 32K | $0.030 | — |
| LFM2-24B-A2B | liquid | — | — | 130 | — | 128K | $0.030 | $0.12 |
| Aion-2.0 | aion labs | — | — | 131 | — | 131K | $0.80 | $1.60 |
| GPT-5 Chat MM | openai | — | — | 132 | — | 128K | $1.25 | $10.0 |
| gpt-oss-safeguard-20b | openai | — | — | 133 | — | 131K | $0.075 | $0.30 |
| Qwen3.6 Max Preview | qwen | — | — | 133 | — | 262K | $1.04 | $6.24 |
| Qwen2.5 7B Instruct | qwen | — | — | 134 | — | 131K | $0.040 | $0.100 |
| GPT-4o-mini (2024-07-18) MM | openai | — | — | 135 | — | 128K | $0.15 | $0.60 |
| Qwen3.5-122B-A10B MM | qwen | — | — | 135 | — | 262K | $0.26 | $2.08 |
| Qwen3 235B A22B Thinking 2507 | qwen | — | — | 136 | — | 262K | $0.15 | $1.50 |
| Claude Opus 4.7 (Fast) MM | anthropic | — | — | 137 | — | 1.0M | $30.0 | $150.0 |
| Qwen3 Coder 30B A3B Instruct OS | qwen | — | — | 138 | — | 160K | $0.070 | $0.27 |
| Ministral 3 14B 2512 MM | mistralai | — | — | 139 | — | 262K | $0.20 | $0.20 |
| Qwen3 14B | qwen | — | — | 139 | — | 132K | $0.100 | $0.24 |
| Nova 2 Lite MM | amazon | — | — | 142 | — | 1.0M | $0.30 | $2.50 |
| Mistral Small 3 | mistralai | — | — | 143 | — | 33K | $0.050 | $0.080 |
| MiniMax M2 | minimax | — | — | 144 | — | 205K | $0.26 | $1.00 |
| Nano Banana 2 (Gemini 3.1 Flash Image Preview) MM | — | — | 144 | — | 131K | $0.50 | $3.00 | |
| Mercury 2 | inception | — | — | 146 | 33 | 128K | $0.25 | $0.75 |
| R1 | deepseek | — | — | 146 | — | 164K | $0.70 | $2.50 |
| Embed V1 0.6B MM | perplexity | — | — | 147 | — | 32K | $0.004 | — |
| Qwen2.5 72B Instruct | qwen | — | — | 148 | — | 131K | $0.36 | $0.40 |
| GPT-5.1-Codex MM | openai | — | — | 149 | — | 400K | $1.25 | $10.0 |
| GPT-5.2 Chat MM | openai | — | — | 152 | — | 128K | $1.75 | $14.0 |
| GPT-5.1-Codex-Mini MM | openai | — | — | 153 | — | 400K | $0.25 | $2.00 |
| GPT-5.5 Pro MM | openai | — | — | 154 | — | 1.1M | $30.0 | $180.0 |
| Kimi K2 0711 | moonshotai | — | — | 154 | — | 131K | $0.57 | $2.30 |
| GLM 4 32B | z ai | — | — | 156 | — | 128K | $0.100 | $0.100 |
| Qwen3 30B A3B | qwen | — | — | 157 | — | 131K | $0.090 | $0.45 |
| Kimi K2 Thinking | moonshotai | — | — | 158 | — | 262K | $0.60 | $2.50 |
| Gemma 3 4B MM | — | — | 159 | — | 131K | $0.040 | $0.080 | |
| GPT-5.4 Image 2 MM | openai | — | — | 160 | — | 272K | $8.00 | $15.0 |
| Nova Micro 1.0 | amazon | — | — | 162 | — | 128K | $0.035 | $0.14 |
| Qwen3.5 Plus 2026-04-20 MM | qwen | — | — | 163 | — | 1.0M | $0.30 | $1.80 |
| Gemini Embedding 2 Preview MM | — | — | 165 | — | 8K | $0.20 | — | |
| Devstral Small 1.1 MM | mistralai | — | — | 166 | — | 131K | $0.100 | $0.30 |
| Ministral 3 3B 2512 MM | mistralai | — | — | 167 | — | 131K | $0.100 | $0.100 |
| Qwen3 235B A22B | qwen | — | — | 167 | — | 131K | $0.45 | $1.82 |
| KAT-Coder-Pro V2 | kwaipilot | — | — | 169 | — | 256K | $0.30 | $1.20 |
| Claude 3 Haiku MM | anthropic | — | — | 170 | — | 200K | $0.25 | $1.25 |
| Mistral Medium 3.1 MM | mistralai | — | — | 172 | — | 131K | $0.40 | $2.00 |
| Trinity Large Thinking | arcee ai | — | — | 172 | 32 | 262K | $0.22 | $0.85 |
| Qwen3 Coder Plus | qwen | — | — | 173 | — | 1.0M | $0.65 | $3.25 |
| Seedream 4.5 MM | bytedance seed | — | — | 174 | — | 4K | — | — |
| Nano Banana Pro (Gemini 3 Pro Image Preview) MM | — | — | 175 | — | 66K | $2.00 | $12.0 | |
| Llama Nemotron Embed VL 1B V2 (free) MM | nvidia | — | — | 176 | — | 131K | — | — |
| GPT-5.2-Codex MM | openai | — | — | 177 | — | 400K | $1.75 | $14.0 |
| Hermes 4 70B | nousresearch | — | — | 177 | — | 131K | $0.13 | $0.40 |
| Qwen2.5 VL 72B Instruct MM | qwen | — | — | 178 | — | 131K | $0.25 | $0.75 |
| Ling-2.6-1T | inclusionai | — | — | 179 | 34 | 262K | $0.075 | $0.63 |
| GLM 4.5 | z ai | — | — | 180 | — | 131K | $0.60 | $2.20 |
| Qwen-Plus | qwen | — | — | 181 | — | 1.0M | $0.26 | $0.78 |
| Sonar MM | perplexity | — | — | 182 | — | 127K | $1.00 | $1.00 |
| Llama Guard 4 12B MM | meta llama | — | — | 183 | — | 164K | $0.18 | $0.18 |
| o4 Mini MM | openai | — | — | 184 | — | 200K | $1.10 | $4.40 |
| GPT-4o (2024-11-20) MM | openai | — | — | 188 | — | 128K | $2.50 | $10.0 |
| MiniMax M2.1 | minimax | — | — | 189 | — | 205K | $0.29 | $0.95 |
| Grok 4.20 Multi-Agent MM | x ai | — | — | 190 | — | 2.0M | $2.00 | $6.00 |
| Llama 3.2 3B Instruct OS | meta llama | — | — | 190 | — | 131K | $0.051 | $0.34 |
About these rankings
Four data sources, refreshed every 6 hours, zero manual curation.
What data sources power the AI Gear Base LLM rankings?
We aggregate live rankings from four authoritative sources: OpenRouter (real-world token usage across 60+ inference providers), LMSys Chatbot Arena (human-voted ELO ratings), Artificial Analysis (independent quality + speed benchmarks), and the Hugging Face open-source leaderboard. Rankings refresh every 6 hours from all sources, so the data you see is never more than 6 hours stale.
How many LLMs and providers are tracked?
1,069 models from 178 providers — including OpenAI, Anthropic, Google, Meta, Mistral, xAI, DeepSeek, Alibaba/Qwen, Cohere, and the leading open-source labs. Every model carries pricing, context window, modalities, and ranking history so you can compare both proprietary frontier models and open-weights releases on equal footing.
How often are the rankings updated?
Every 6 hours, fully automated. A Cloudflare cron trigger pulls fresh data from the OpenRouter and Hugging Face public APIs, while Apify Actors scrape the LMSys Arena and Artificial Analysis leaderboards on the same cadence and post the results to our ingestion webhook. All snapshots are written to a time-series table so we can show 7-day, 30-day, and all-time trend lines on each model's detail page.
Why do rankings differ between sources?
Each source measures a different thing. LMSys Arena captures human preference (which model do real users like more?), Artificial Analysis measures objective quality + throughput on standardised benchmarks, OpenRouter reflects actual production usage by paying API customers, and Hugging Face tracks community popularity. A truly great model usually ranks highly across all four — which is why we show every score side-by-side in the comparison table instead of inventing a single "AI Gear Base score" that hides the nuance.