LMArena Vision Elo Leaderboard
vision113 models · Updated October 2026
On the LMArena Vision Elo benchmark, Anthropic: Claude Fable 5 ranks #1 with a score of 1324, while Pixtral 12B offers the best score-per-dollar at $0.15/1M output tokens. The full ranking, with cost per million tokens, is below.
💰 Best value
Pixtral 12B — score 1009 at $0.15/1M output tokens
| # | Model | Score | Output / 1M | Context |
|---|---|---|---|---|
| 1 | Anthropic: Claude Fable 5Anthropic | 1324 | $50.00 | 1M |
| 2 | Anthropic: Claude Opus 5.5Anthropic | 1321 | $20.00 | 1M |
| 3 | Anthropic: Claude Opus 5Anthropic | 1319 | $25.00 | 1M |
| 4 | Anthropic: Claude Fable 5.1Anthropic | 1318 | $50.00 | 1M |
| 5 | Anthropic: Claude Opus 4.7Anthropic | 1316 | $25.00 | 1M |
| 6 | Google: Gemini 3.7 FlashGoogle | 1316 | $3.75 | 1M |
| 7 | Anthropic: Claude Opus 4.6Anthropic | 1316 | $25.00 | 1M |
| 8 | Google: Gemini 3.8 FlashGoogle | 1314 | $3.75 | 1M |
| 9 | Qwen3.8 MaxAlibaba | 1314 | $6.00 | 1M |
| 10 | Google: Gemini 3.5 FlashGoogle | 1310 | $9.00 | 1M |
| 11 | Meta: Muse Spark 1.3Meta | 1309 | $4.25 | 1M |
| 12 | Muse SparkMeta AI | 1306 | — | — |
| 13 | Gemini 3 ProGoogle DeepMind | 1305 | — | — |
| 14 | Meta: Muse Spark 1.2Meta | 1305 | $4.25 | 1M |
| 15 | OpenAI: GPT-5.4OpenAI | 1303 | $15.00 | 1.1M |
| 16 | Google: Gemini 3.6 FlashGoogle | 1298 | $3.75 | 1M |
| 17 | OpenAI: GPT-5.5OpenAI | 1297 | $30.00 | 1.1M |
| 18 | Google: Gemini 3.1 Pro PreviewGoogle | 1296 | $12.00 | 1M |
| 19 | Z.ai: GLM 5.3 FlashZ.ai | 1296 | $0.50 | 1M |
| 20 | Anthropic: Claude Opus 4.8Anthropic | 1294 | $25.00 | 1M |
| 21 | Meta: Muse Spark 1.1Meta | 1293 | $4.25 | 1M |
| 22 | Anthropic: Claude Sonnet 5.5Anthropic | 1289 | $10.00 | 1M |
| 23 | OpenAI: GPT-6.1 SolOpenAI | 1288 | $10.00 | 1.1M |
| 24 | SpaceXAI: Grok 4.5SpaceXAI | 1288 | $6.00 | 500K |
| 25 | Gemini 3 FlashOpenCode | 1285 | $3.00 | 1M |
| 26 | MoonshotAI: Kimi K2.6MoonshotAI | 1283 | $3.41 | 262K |
| 27 | Anthropic: Claude Sonnet 4.6Anthropic | 1282 | $15.00 | 1M |
| 28 | OpenAI: GPT-5.6 SolOpenAI | 1281 | $10.00 | 1.1M |
| 29 | OpenAI: GPT-6 AstraOpenAI | 1281 | $50.00 | 1.1M |
| 30 | Qwen3.7 PlusAlibaba | 1279 | $1.60 | 1M |
| 31 | Google: Gemma 4 31BGoogle | 1277 | $0.34 | 262K |
| 32 | DeepSeek: DeepSeek V4.1 FlashDeepSeek | 1277 | $0.89 | 1M |
| 33 | Anthropic: Claude Sonnet 5Anthropic | 1274 | $10.00 | 1M |
| 34 | Qwen: Qwen3.8 27BQwen | 1271 | $2.30 | 1M |
| 35 | OpenAI: GPT-5.6 TerraOpenAI | 1271 | $12.00 | 1.1M |
| 36 | MoonshotAI: Kimi K2.5MoonshotAI | 1269 | $2.50 | 262K |
| 37 | Google: Gemini 3.5 Flash LiteGoogle | 1268 | $2.50 | 1M |
| 38 | gpt-5.2-chat-latest302.AI | 1268 | $14.00 | 128K |
| 39 | StepFun: Step 5 PreviewStepFun | 1267 | $2.70 | 1M |
| 40 | Z.ai: GLM 5V TurboZ.ai | 1264 | $4.00 | 203K |
| 41 | Xiaomi: MiMo-V2.6-ProXiaomi | 1264 | $0.87 | 1.1M |
| 42 | Google: Gemini 2.5 ProGoogle | 1263 | $10.00 | 1M |
| 43 | SpaceXAI: Grok 4.6SpaceXAI | 1263 | $6.00 | 500K |
| 44 | grok-4.20-beta-0309-reasoning302.AI | 1263 | $6.00 | 2M |
| 45 | Qwen3.5 397B-A17BAlibaba | 1263 | $3.60 | 262K |
| 46 | Google: Gemma 4 26B A4B Google | 1260 | $0.26 | 262K |
| 47 | Xiaomi: MiMo-V2.6-FlashXiaomi | 1259 | $0.28 | 1.1M |
| 48 | grok-4.20-multi-agent-beta-0309302.AI | 1259 | $6.00 | 2M |
| 49 | OpenAI: GPT-5.6 LunaOpenAI | 1258 | $1.20 | 1.1M |
| 50 | MiniMax: MiniMax M3MiniMax | 1253 | $1.20 | 1M |
| 51 | gemini-2.5-flash-preview-09-2025302.AI | 1253 | $2.50 | 1M |
| 52 | GPT-5.5 InstantOpenAI | 1250 | — | — |
| 53 | OpenAI: GPT-5.1OpenAI | 1250 | $10.00 | 400K |
| 54 | Xiaomi: MiMo-V2.5Xiaomi | 1247 | $0.28 | 1.1M |
| 55 | Qwen: Qwen3 VL 235B A22B InstructQwen | 1247 | $1.90 | 262K |
| 56 | OpenAI: GPT-5.2OpenAI | 1246 | $14.00 | 400K |
| 57 | OpenAI: GPT-5.4 MiniOpenAI | 1245 | $4.50 | 400K |
| 58 | OpenAI: GPT-6 SolOpenAI | 1245 | $10.00 | 1.1M |
| 59 | Qwen3.5 122B-A10BAlibaba | 1245 | $3.20 | 262K |
| 60 | chatgpt-4o-latest302.AI | 1243 | $15.00 | 128K |
| 61 | Qwen3.5 27BAlibaba | 1241 | $2.40 | 262K |
| 62 | Google: Gemini 3.1 Flash Lite PreviewGoogle | 1240 | $1.50 | 1M |
| 63 | Google: Gemini 2.5 FlashGoogle | 1235 | $2.50 | 1M |
| 64 | Thinking Machines: Inkling SmallThinking Machines | 1235 | $1.20 | 524K |
| 65 | OpenAI: GPT-5 ChatOpenAI | 1232 | $10.00 | 128K |
| 66 | SpaceXAI: Grok 4.3SpaceXAI | 1229 | $2.50 | 1M |
| 67 | SpaceXAI: Grok 4.7SpaceXAI | 1228 | $6.00 | 500K |
| 68 | MiMo-V2-OmniXiaomi | 1228 | $0.28 | 262K |
| 69 | Mistral: Mistral Medium 3.5Mistral | 1223 | $7.50 | 262K |
| 70 | Mistral: Mistral Large 3 2512Mistral | 1221 | $1.50 | 262K |
| 71 | OpenAI: GPT-6 LunaOpenAI | 1217 | $0.50 | 1.1M |
| 72 | OpenAI: o3OpenAI | 1214 | $8.00 | 200K |
| 73 | Qwen-VL MaxAlibaba | 1213 | $3.20 | 131K |
| 74 | OpenAI: GPT-4.1OpenAI | 1211 | $8.00 | 1M |
| 75 | Grok 4xAI | 1210 | — | — |
| 76 | OpenAI: GPT-5OpenAI | 1208 | $10.00 | 400K |
| 77 | Qwen3-VL 235B-A22BAlibaba | 1207 | $2.80 | 131K |
| 78 | OpenAI: GPT-5 MiniOpenAI | 1202 | $2.00 | 400K |
| 79 | Grok 4.1 FastxAI | 1201 | — | — |
| 80 | OpenAI: GPT-5.4 NanoOpenAI | 1196 | $1.25 | 400K |
| 81 | OpenAI: o4 MiniOpenAI | 1194 | $4.40 | 200K |
| 82 | Anthropic: Claude Opus 4Anthropic | 1192 | $75.00 | 200K |
| 83 | Anthropic: Claude Sonnet 4Anthropic | 1191 | $15.00 | 200K |
| 84 | gemini-2.5-flash-lite-preview-06-17Jiekou.AI | 1187 | $0.36 | 1M |
| 85 | OpenAI: GPT-4.1 MiniOpenAI | 1181 | $1.60 | 1M |
| 86 | Step-3NanoGPT | 1176 | $0.65 | 66K |
| 87 | Mistral Medium (latest)Mistral | 1172 | $7.50 | 262K |
| 88 | Claude 3.7 SonnetAnthropic | 1169 | — | — |
| 89 | OpenAI: o1OpenAI | 1168 | $60.00 | 200K |
| 90 | Gemma 3 27B ITCortecs | 1164 | $0.30 | 131K |
| 91 | Z.ai: GLM 4.6VZ.ai | 1164 | $0.90 | 131K |
| 92 | OpenAI: GPT-5 NanoOpenAI | 1159 | $0.40 | 400K |
| 93 | Z.ai: GLM 4.5VZ.ai | 1154 | $1.80 | 66K |
| 94 | Qwen2.5 VL 32B InstructLLM Gateway | 1152 | $4.20 | 131K |
| 95 | Mistral Small (latest)Mistral | 1142 | $0.60 | 262K |
| 96 | Llama 4 Maverick 17B 128E InstructVertex | 1142 | $1.15 | 524K |
| 97 | OpenAI: GPT-4oOpenAI | 1137 | $10.00 | 128K |
| 98 | Mistral Small 31 24b InstructNanoGPT | 1136 | $0.30 | 128K |
| 99 | Claude 3.5 SonnetAnthropic | 1125 | — | — |
| 100 | Llama 4 Scout 17B 16E InstructAzure | 1118 | $0.78 | 128K |
| 101 | Qwen2.5-VL 72B InstructAlibaba | 1107 | $8.40 | 131K |
| 102 | Anthropic: Claude 3.5 HaikuHelicone | 1092 | $4.00 | 200K |
| 103 | OpenAI: GPT-4 TurboOpenAI | 1090 | $30.00 | 128K |
| 104 | Pixtral Large (latest)Mistral | 1089 | $6.00 | 128K |
| 105 | OpenAI: GPT-4o-miniOpenAI | 1066 | $0.60 | 128K |
| 106 | OpenAI: GPT-4.1 NanoOpenAI | 1063 | $0.40 | 1M |
| 107 | Claude 3 OpusAnthropic | 1023 | — | — |
| 108 | Pixtral 12BMistral | 1009 | $0.15 | 128K |
| 109 | Amazon: Nova Lite 1.0Amazon | 990 | $0.24 | 300K |
| 110 | Claude 3 SonnetAnthropic | 984 | — | — |
| 111 | Amazon: Nova Pro 1.0Amazon | 980 | $3.20 | 300K |
| 112 | Anthropic: Claude 3 HaikuAnthropic | 950 | $1.25 | 200K |
| 113 | Phi-3.5-vision instruct (128k)GitHub Models | 851 | $0.00 | 128K |
What does LMArena Vision Elo test?
Overview This dataset contains ALL in-the-wild conversation crowdsourced from Search Arena between March 18, 2025 and May 8, 2025. It includes 24,069 multi-turn conversations with search-LLMs across diverse intents, languages, and topics—alongs...
Frequently asked questions
Pricing is indicative — confirm with the provider before production use. Updated October 2026.