Best LLM for vision
Vision covers reading a document, following a diagram, and describing what is in a frame.
Aldena runs these models inside your team's projects. See what each one costs.
| rank | model | vendor | composite | pricein / out | benchmarks | elo | elo | elo | elo | elo | elo |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | Claude Fable 5 (high)anthropic/claude-fable-5:high | Anthropic | 83.8 | $10.00 / $50.00 | 4 of 6 benchmarks | 1309 | 1327 | 1350 | 1353 | ||
| 2 | Claude Opus 5.5 (high)anthropic/claude-opus-5.5:high | Anthropic | 82.1 | $4.00 / $20.00 | 4 of 6 benchmarks | 1295 | 1312 | 1338 | 1347 | ||
| 3 | Gemini 3.8 Flash (high)google/gemini-3.8-flash:high | 81.9 | $0.75 / $3.75 | 4 of 6 benchmarks | 1293 | 1315 | 1341 | 1339 | |||
| 4 | Claude Opus 4.7 (high)anthropic/claude-opus-4.7:high | Anthropic | 81.8 | $5.00 / $25.00 | 5 of 6 benchmarks | 1298 | 1311 | 1334 | 1331 | 1246 | |
| 5 | Qwen3.8 Maxqwen/qwen3.8-max | Qwen | 81.2 | — | 4 of 6 benchmarks | 1299 | 1316 | 1322 | 1334 | ||
| 6 | Claude Opus 4.6 (high)anthropic/claude-opus-4.6:high | Anthropic | 79.6 | $5.00 / $25.00 | 5 of 6 benchmarks | 1299 | 1313 | 1324 | 1324 | 1237 | |
| 7 | GPT-6.1 Sol (max)openai/gpt-6.1-sol:max | OpenAI | 79.2 | $2.00 / $10.00 | 4 of 6 benchmarks | 1292 | 1299 | 1327 | 1354 | ||
| 8 | Claude Opus 4.6anthropic/claude-opus-4.6 | Anthropic | 78.4 | $5.00 / $25.00 | 5 of 6 benchmarks | 1293 | 1311 | 1330 | 1311 | 1239 | |
| 9 | Gemini 3 Progoogle/gemini-3-pro | 78.3 | — | 6 of 6 benchmarks | 1289 | 1302 | 1302 | 1308 | 1290 | 1277 | |
| 10 | Claude Fable 5.1 (max)anthropic/claude-fable-5.1:max | Anthropic | 78.2 | $10.00 / $50.00 | 4 of 6 benchmarks | 1289 | 1304 | 1328 | 1328 | ||
| 11 | Gemini 3.1 Pro Previewgoogle/gemini-3.1-pro-preview | 77.2 | $2.00 / $12.00 | 6 of 6 benchmarks | 1280 | 1295 | 1309 | 1311 | 1284 | 1280 | |
| 12 | Gemini 3.7 Flash (high)google/gemini-3.7-flash:high | 77.2 | $0.75 / $3.75 | 4 of 6 benchmarks | 1294 | 1308 | 1320 | 1313 | |||
| 13 | Claude Opus 4.8 (high)anthropic/claude-opus-4.8:high | Anthropic | 76.3 | $5.00 / $25.00 | 4 of 6 benchmarks | 1286 | 1299 | 1317 | 1331 | ||
| 14 | Claude Opus 5 (high)anthropic/claude-opus-5:high | Anthropic | 76.3 | $5.00 / $25.00 | 4 of 6 benchmarks | 1285 | 1300 | 1326 | 1321 | ||
| 15 | Muse Spark 1.3 (max)meta/muse-spark-1.3:max | Meta | 75.1 | $1.25 / $4.25 | 4 of 6 benchmarks | 1291 | 1308 | 1329 | 1301 | ||
| 16 | GPT-6 Astra (max)openai/gpt-6-astra:max | OpenAI | 75.0 | $10.00 / $50.00 | 4 of 6 benchmarks | 1284 | 1298 | 1311 | 1341 | ||
| 17 | Muse Spark 1.2 (xhigh)meta/muse-spark-1.2:xhigh | Meta | 74.9 | $1.25 / $4.25 | 3 of 6 benchmarks | 1294 | 1306 | 1316 | |||
| 18 | GPT-5.6 Sol (xhigh)openai/gpt-5.6-sol:xhigh | OpenAI | 74.8 | $4.00 / $20.00 | 4 of 6 benchmarks | 1288 | 1298 | 1318 | 1316 | ||
| 19 | Muse Spark 1.3 (xhigh)meta/muse-spark-1.3:xhigh | Meta | 74.3 | $1.25 / $4.25 | 4 of 6 benchmarks | 1288 | 1301 | 1315 | 1311 | ||
| 20 | GPT-5.5 (high)openai/gpt-5.5:high | OpenAI | 74.1 | $5.00 / $30.00 | 5 of 6 benchmarks | 1283 | 1298 | 1316 | 1335 | 1223 | |
| 21 | Claude Opus 4.7anthropic/claude-opus-4.7 | Anthropic | 73.4 | $5.00 / $25.00 | 5 of 6 benchmarks | 1300 | 1315 | 1330 | 1325 | 1183 | |
| 22 | Muse Sparkmeta/muse-spark | Meta | 73.0 | — | 4 of 6 benchmarks | 1294 | 1303 | 1312 | 1299 | ||
| 23 | Gemini 3 Flash Previewgoogle/gemini-3-flash-preview | 72.4 | $0.50 / $3.00 | 6 of 6 benchmarks | 1272 | 1285 | 1293 | 1309 | 1292 | 1238 | |
| 24 | Gemini 3.5 Flash (medium)google/gemini-3.5-flash:medium | 72.2 | $1.50 / $9.00 | 4 of 6 benchmarks | 1281 | 1298 | 1310 | 1314 | |||
| 25 | GPT-5.5openai/gpt-5.5 | OpenAI | 71.9 | $5.00 / $30.00 | 5 of 6 benchmarks | 1285 | 1296 | 1321 | 1330 | 1213 | |
| 26 | Gemini 3.6 Flash (high)google/gemini-3.6-flash:high | 71.2 | $0.75 / $3.75 | 4 of 6 benchmarks | 1280 | 1298 | 1312 | 1308 | |||
| 27 | Claude Opus 4.8anthropic/claude-opus-4.8 | Anthropic | 70.3 | $5.00 / $25.00 | 4 of 6 benchmarks | 1279 | 1293 | 1306 | 1321 | ||
| 28 | GLM 5.3 Flashz-ai/glm-5.3-flash | Z.ai | 69.6 | $0.15 / $0.50 | 5 of 6 benchmarks | 1272 | 1287 | 1310 | 1312 | 1237 | |
| 29 | GPT-5.4 (high)openai/gpt-5.4:high | OpenAI | 69.4 | $2.50 / $15.00 | 5 of 6 benchmarks | 1282 | 1297 | 1319 | 1330 | 1193 | |
| 30 | Gemini 3.5 Flash (high)google/gemini-3.5-flash:high | 68.9 | $1.50 / $9.00 | 4 of 6 benchmarks | 1281 | 1291 | 1307 | 1307 | |||
| 31 | GPT-5.2 Chatopenai/gpt-5.2-chat | OpenAI | 68.7 | — | 5 of 6 benchmarks | 1278 | 1288 | 1307 | 1293 | 1241 | |
| 32 | Grok 4.5x-ai/grok-4.5 | SpaceXAI | 68.1 | $2.00 / $6.00 | 4 of 6 benchmarks | 1280 | 1290 | 1307 | 1305 | ||
| 33 | Claude Sonnet 5.5 (xhigh)anthropic/claude-sonnet-5.5:xhigh | Anthropic | 67.5 | $2.00 / $10.00 | 4 of 6 benchmarks | 1263 | 1281 | 1306 | 1319 | ||
| 34 | GPT-5.6 Terra (xhigh)openai/gpt-5.6-terra:xhigh | OpenAI | 66.3 | $2.00 / $12.00 | 4 of 6 benchmarks | 1266 | 1277 | 1296 | 1318 | ||
| 35 | Gemini 3 Flash Preview (minimal)google/gemini-3-flash-preview:minimal | 65.8 | $0.50 / $3.00 | 6 of 6 benchmarks | 1260 | 1272 | 1285 | 1297 | 1274 | 1212 | |
| 36 | GPT-5.4openai/gpt-5.4 | OpenAI | 65.7 | $2.50 / $15.00 | 5 of 6 benchmarks | 1278 | 1290 | 1313 | 1304 | 1213 | |
| 37 | Muse Spark 1.1meta/muse-spark-1.1 | Meta | 65.3 | $1.25 / $4.25 | 4 of 6 benchmarks | 1279 | 1294 | 1305 | 1285 | ||
| 38 | GPT 5.5 Instantopenai/gpt-5.5-instant | OpenAI | 64.4 | — | 4 of 6 benchmarks | 1277 | 1287 | 1310 | 1279 | ||
| 39 | GPT-6 Sol (max)openai/gpt-6-sol:max | OpenAI | 64.3 | $2.00 / $10.00 | 4 of 6 benchmarks | 1258 | 1278 | 1297 | 1310 | ||
| 40 | Kimi K2.6moonshotai/kimi-k2.6 | MoonshotAI | 63.2 | $0.65 / $3.41 | 4 of 6 benchmarks | 1265 | 1280 | 1293 | 1303 | ||
| 41 | DeepSeek V4.1 Flash (max)deepseek/deepseek-v4.1-flash:max | DeepSeek | 63.1 | $0.051 / $0.567 | 4 of 6 benchmarks | 1263 | 1274 | 1296 | 1307 | ||
| 42 | Claude Sonnet 5 (high)anthropic/claude-sonnet-5:high | Anthropic | 62.4 | $2.00 / $10.00 | 4 of 6 benchmarks | 1261 | 1279 | 1292 | 1303 | ||
| 43 | Gemini 2.5 Progoogle/gemini-2.5-pro | 61.6 | $1.25 / $10.00 | 6 of 6 benchmarks | 1247 | 1259 | 1267 | 1273 | 1251 | 1251 | |
| 44 | Gemini 3.5 Flash Litegoogle/gemini-3.5-flash-lite | 61.5 | $0.30 / $2.50 | 4 of 6 benchmarks | 1266 | 1280 | 1287 | 1288 | |||
| 45 | GPT-5.1 (high)openai/gpt-5.1:high | OpenAI | 61.4 | $1.25 / $10.00 | 6 of 6 benchmarks | 1251 | 1262 | 1277 | 1285 | 1229 | 1249 |
| 46 | Qwen3.7 Plusqwen/qwen3.7-plus | Qwen | 61.2 | $0.32 / $1.28 | 4 of 6 benchmarks | 1262 | 1280 | 1288 | 1288 | ||
| 47 | Grok 4.6 (high)x-ai/grok-4.6:high | SpaceXAI | 61.0 | $2.00 / $6.00 | 4 of 6 benchmarks | 1263 | 1275 | 1295 | 1288 | ||
| 48 | Kimi K2.5 (thinking)moonshotai/kimi-k2.5:thinking | MoonshotAI | 59.2 | $0.532 / $3.325 | 6 of 6 benchmarks | 1252 | 1266 | 1280 | 1289 | 1236 | 1193 |
| 49 | Grok 4.7 (xhigh)x-ai/grok-4.7:xhigh | SpaceXAI | 59.2 | $2.00 / $6.00 | 4 of 6 benchmarks | 1249 | 1262 | 1284 | 1307 | ||
| 50 | Grok 4.20 Beta 0309 (reasoning)x-ai/grok-4.20-beta-0309:reasoning | SpaceXAI | 58.9 | — | 5 of 6 benchmarks | 1256 | 1267 | 1275 | 1266 | 1240 | |
| 51 | Claude Sonnet 4.6anthropic/claude-sonnet-4.6 | Anthropic | 58.8 | $3.00 / $15.00 | 5 of 6 benchmarks | 1275 | 1290 | 1305 | 1302 | 1168 | |
| 52 | Dola Seed 2.0 Probytedance/dola-seed-2.0-pro | ByteDance | 57.9 | — | 4 of 6 benchmarks | 1256 | 1270 | 1288 | 1277 | ||
| 53 | GPT-5.6 Luna (xhigh)openai/gpt-5.6-luna:xhigh | OpenAI | 57.6 | $0.20 / $1.20 | 4 of 6 benchmarks | 1255 | 1265 | 1278 | 1289 | ||
| 54 | GPT-5.2 (high)openai/gpt-5.2:high | OpenAI | 57.3 | $1.75 / $14.00 | 6 of 6 benchmarks | 1243 | 1260 | 1276 | 1281 | 1202 | 1261 |
| 55 | Gemini 3.1 Flash Lite Previewgoogle/gemini-3.1-flash-lite-preview | 57.2 | $0.25 / $1.50 | 6 of 6 benchmarks | 1235 | 1247 | 1252 | 1271 | 1269 | 1218 | |
| 56 | GPT-5.4 Mini (high)openai/gpt-5.4-mini:high | OpenAI | 55.7 | $0.75 / $4.50 | 5 of 6 benchmarks | 1250 | 1263 | 1283 | 1291 | 1204 | |
| 57 | Grok 4.20 Multi Agent Beta 0309x-ai/grok-4.20-multi-agent-beta-0309 | SpaceXAI | 55.3 | — | 5 of 6 benchmarks | 1251 | 1260 | 1267 | 1274 | 1228 | |
| 58 | Step 5 Previewstepfun/step-5-preview | StepFun | 55.2 | $1.00 / $2.70 | 4 of 6 benchmarks | 1246 | 1266 | 1288 | 1269 | ||
| 59 | Gemma 4 31Bgoogle/gemma-4-31b-it | 54.5 | $0.20 / $0.40 | 6 of 6 benchmarks | 1265 | 1277 | 1287 | 1304 | 1198 | 1161 | |
| 60 | MiMo-V2.6-Flashxiaomi/mimo-v2.6-flash | Xiaomi | 54.0 | $0.14 / $0.28 | 4 of 6 benchmarks | 1248 | 1265 | 1265 | 1283 | ||
| 61 | Qwen3.8 27Bqwen/qwen3.8-27b | Qwen | 53.1 | $0.15 / $1.875 | 4 of 6 benchmarks | 1244 | 1262 | 1267 | 1275 | ||
| 62 | Qwen3.5 397B A17Bqwen/qwen3.5-397b-a17b | Qwen | 52.1 | $0.60 / $3.60 | 6 of 6 benchmarks | 1246 | 1260 | 1273 | 1287 | 1220 | 1171 |
| 63 | GPT-5 (high)openai/gpt-5:high | OpenAI | 51.3 | $1.25 / $10.00 | 6 of 6 benchmarks | 1211 | 1232 | 1251 | 1259 | 1261 | 1195 |
| 64 | MiMo-V2.6-Proxiaomi/mimo-v2.6-pro | Xiaomi | 51.0 | $0.435 / $0.87 | 4 of 6 benchmarks | 1247 | 1259 | 1272 | 1256 | ||
| 65 | Chatgpt 4oopenai/chatgpt-4o | OpenAI | 50.6 | — | 6 of 6 benchmarks | 1238 | 1246 | 1260 | 1245 | 1218 | 1215 |
| 66 | Gemma 4 26B A4B google/gemma-4-26b-a4b-it | 50.1 | $0.0765 / $0.255 | 6 of 6 benchmarks | 1242 | 1257 | 1261 | 1285 | 1159 | 1229 | |
| 67 | GPT 4.5 Previewopenai/gpt-4.5-preview | OpenAI | 49.9 | — | 1 of 6 benchmarks | 1225 | |||||
| 68 | Grok 4.3x-ai/grok-4.3 | SpaceXAI | 49.7 | $1.25 / $2.50 | 5 of 6 benchmarks | 1242 | 1251 | 1267 | 1255 | 1214 | |
| 69 | GPT-6 Luna (max)openai/gpt-6-luna:max | OpenAI | 49.6 | $0.10 / $0.50 | 4 of 6 benchmarks | 1236 | 1248 | 1267 | 1271 | ||
| 70 | GPT 5 Chatopenai/gpt-5-chat | OpenAI | 47.7 | — | 6 of 6 benchmarks | 1225 | 1245 | 1258 | 1275 | 1189 | 1195 |
| 71 | Mistral Large 4mistralai/mistral-large-4-0 | Mistral | 46.5 | $0.70 / $2.10 | 4 of 6 benchmarks | 1230 | 1242 | 1265 | 1256 | ||
| 72 | o1openai/o1 | OpenAI | 45.4 | — | 1 of 6 benchmarks | 1193 | |||||
| 73 | Gemini 2.5 Flashgoogle/gemini-2.5-flash | 45.0 | $0.30 / $2.50 | 6 of 6 benchmarks | 1215 | 1223 | 1233 | 1239 | 1221 | 1210 | |
| 74 | o3openai/o3 | OpenAI | 44.9 | $2.00 / $8.00 | 6 of 6 benchmarks | 1215 | 1226 | 1233 | 1249 | 1231 | 1184 |
| 75 | Kimi K2.5 Instantmoonshotai/kimi-k2.5-instant | Moonshot AI | 44.7 | — | 4 of 6 benchmarks | 1236 | 1244 | 1251 | 1248 | ||
| 76 | Qwen3 VL 235B A22B Instructqwen/qwen3-vl-235b-a22b-instruct | Qwen | 43.7 | $0.21 / $1.90 | 6 of 6 benchmarks | 1215 | 1227 | 1238 | 1258 | 1176 | 1212 |
| 77 | GPT-5.1openai/gpt-5.1 | OpenAI | 43.7 | $1.25 / $10.00 | 6 of 6 benchmarks | 1240 | 1252 | 1262 | 1257 | 1188 | 1165 |
| 78 | Qwen3.5-122B-A10Bqwen/qwen3.5-122b-a10b | Qwen | 43.6 | $0.26 / $2.08 | 4 of 6 benchmarks | 1226 | 1239 | 1244 | 1254 | ||
| 79 | MiMo-V2.5xiaomi/mimo-v2.5 | Xiaomi | 43.5 | $0.40 / $2.00 | 5 of 6 benchmarks | 1234 | 1248 | 1267 | 1254 | 1169 | |
| 80 | Gemini 1.5 Pro 002google/gemini-1.5-pro-002 | 43.2 | — | 1 of 6 benchmarks | 1180 | ||||||
| 81 | Mimo v2 Omnixiaomi/mimo-v2-omni | Xiaomi | 42.7 | — | 4 of 6 benchmarks | 1216 | 1231 | 1255 | 1248 | ||
| 82 | GPT-4.1openai/gpt-4.1 | OpenAI | 42.3 | $2.00 / $8.00 | 6 of 6 benchmarks | 1214 | 1227 | 1232 | 1252 | 1184 | 1211 |
| 83 | GPT-4o (2024-05-13)openai/gpt-4o-2024-05-13 | OpenAI | 42.3 | $5.00 / $15.00 | 1 of 6 benchmarks | 1162 | |||||
| 84 | GLM 5V Turboz-ai/glm-5v-turbo | Z.ai | 42.1 | $1.20 / $4.00 | 6 of 6 benchmarks | 1232 | 1245 | 1264 | 1252 | 1174 | 1173 |
| 85 | o4 Miniopenai/o4-mini | OpenAI | 41.9 | — | 6 of 6 benchmarks | 1201 | 1211 | 1219 | 1254 | 1221 | 1184 |
| 86 | MiniMax M3minimax/minimax-m3 | MiniMax | 41.8 | $0.30 / $1.20 | 5 of 6 benchmarks | 1232 | 1244 | 1260 | 1258 | 1160 | |
| 87 | GPT-5.2openai/gpt-5.2 | OpenAI | 40.5 | $1.75 / $14.00 | 6 of 6 benchmarks | 1228 | 1238 | 1257 | 1264 | 1172 | 1166 |
| 88 | Ernie 5.0 Preview 1220baidu/ernie-5.0-preview-1220 | Baidu | 39.9 | — | 4 of 6 benchmarks | 1219 | 1230 | 1222 | 1243 | ||
| 89 | Claude Sonnet 4 (thinking 32K)anthropic/claude-sonnet-4:thinking-32k | Anthropic | 39.3 | — | 3 of 6 benchmarks | 1206 | 1216 | 1228 | |||
| 90 | Gemini 1.5 Flash 002google/gemini-1.5-flash-002 | 39.2 | — | 1 of 6 benchmarks | 1142 | ||||||
| 91 | Inkling Smallthinkingmachines/inkling-small | Thinking Machines | 38.8 | $0.45 / $1.20 | 4 of 6 benchmarks | 1208 | 1222 | 1240 | 1239 | ||
| 92 | Gemini 2.0 Flash Lite Preview 02 05google/gemini-2.0-flash-lite-preview-02-05 | 38.8 | — | 1 of 6 benchmarks | 1136 | ||||||
| 93 | Qwen VL (max)qwen/qwen-vl:max | Qwen | 38.6 | — | 3 of 6 benchmarks | 1185 | 1213 | 1248 | |||
| 94 | Nemotron 3.5 Super VLnvidia/nemotron-3.5-super-vl | NVIDIA | 38.6 | $0.15 / $0.47 | 4 of 6 benchmarks | 1212 | 1224 | 1239 | 1233 | ||
| 95 | Qwen2.5 VL 72B Instructqwen/qwen2.5-vl-72b-instruct | Qwen | 37.8 | — | 1 of 6 benchmarks | 1121 | |||||
| 96 | Gemini 1.5 Pro 001google/gemini-1.5-pro-001 | 37.6 | — | 1 of 6 benchmarks | 1120 | ||||||
| 97 | Qwen2.5 VL 32B Instructqwen/qwen2.5-vl-32b-instruct | Qwen | 37.3 | — | 1 of 6 benchmarks | 1118 | |||||
| 98 | GPT-4o (2024-08-06)openai/gpt-4o-2024-08-06 | OpenAI | 37.1 | $2.50 / $10.00 | 1 of 6 benchmarks | 1118 | |||||
| 99 | GPT-4 Turboopenai/gpt-4-turbo | OpenAI | 36.6 | — | 1 of 6 benchmarks | 1112 | |||||
| 100 | GPT-4.1 Miniopenai/gpt-4.1-mini | OpenAI | 36.5 | $0.40 / $1.60 | 6 of 6 benchmarks | 1201 | 1207 | 1208 | 1221 | 1197 | 1193 |
| 101 | Mistral Medium 3.5mistralai/mistral-medium-3-5 | Mistral | 36.5 | $1.50 / $7.50 | 4 of 6 benchmarks | 1200 | 1215 | 1243 | 1216 | ||
| 102 | GPT-4o-mini (2024-07-18)openai/gpt-4o-mini-2024-07-18 | OpenAI | 36.4 | $0.15 / $0.60 | 1 of 6 benchmarks | 1097 | |||||
| 103 | GPT-4.1 Nanoopenai/gpt-4.1-nano | OpenAI | 36.1 | $0.10 / $0.40 | 1 of 6 benchmarks | 1089 | |||||
| 104 | GPT-5 Mini (high)openai/gpt-5-mini:high | OpenAI | 36.1 | $0.25 / $2.00 | 6 of 6 benchmarks | 1185 | 1200 | 1213 | 1238 | 1212 | 1173 |
| 105 | Gemini 1.5 Flash 8B 001google/gemini-1.5-flash-8b-001 | 35.9 | — | 1 of 6 benchmarks | 1071 | ||||||
| 106 | Claude 3 Opusanthropic/claude-3-opus | Anthropic | 35.7 | — | 1 of 6 benchmarks | 1062 | |||||
| 107 | Grok 4.1 Fast (reasoning)x-ai/grok-4-1-fast:reasoning | SpaceXAI | 35.4 | — | 5 of 6 benchmarks | 1196 | 1194 | 1218 | 1171 | 1220 | |
| 108 | Gemini 1.5 Flash 001google/gemini-1.5-flash-001 | 35.4 | — | 1 of 6 benchmarks | 1059 | ||||||
| 109 | Amazon Nova Lite v1.0amazon/amazon-nova-lite-v1.0 | Amazon | 35.2 | — | 1 of 6 benchmarks | 1019 | |||||
| 110 | Mistral Large 3mistralai/mistral-large-3 | Mistral AI | 35.1 | — | 4 of 6 benchmarks | 1200 | 1214 | 1233 | 1203 | ||
| 111 | Amazon Nova Pro v1.0amazon/amazon-nova-pro-v1.0 | Amazon | 35.0 | — | 1 of 6 benchmarks | 1019 | |||||
| 112 | Claude Opus 4 (thinking 16K)anthropic/claude-opus-4:thinking-16k | Anthropic | 34.9 | — | 4 of 6 benchmarks | 1208 | 1219 | 1209 | 1213 | ||
| 113 | Claude 3 Sonnetanthropic/claude-3-sonnet | Anthropic | 34.7 | — | 1 of 6 benchmarks | 1017 | |||||
| 114 | Claude 3 Haikuanthropic/claude-3-haiku | Anthropic | 34.5 | — | 1 of 6 benchmarks | 1000 | |||||
| 115 | Grok 4 0709x-ai/grok-4-0709 | SpaceXAI | 34.3 | — | 6 of 6 benchmarks | 1184 | 1176 | 1192 | 1171 | 1236 | 1171 |
| 116 | Claude Opus 4anthropic/claude-opus-4 | Anthropic | 34.0 | — | 4 of 6 benchmarks | 1189 | 1198 | 1206 | 1243 | ||
| 117 | Claude 3.7 Sonnet (thinking 32K)anthropic/claude-3-7-sonnet:thinking-32k | Anthropic | 33.2 | — | 4 of 6 benchmarks | 1193 | 1210 | 1213 | 1214 | ||
| 118 | Qwen3 VL 235B A22B Thinkingqwen/qwen3-vl-235b-a22b-thinking | Qwen | 32.6 | — | 4 of 6 benchmarks | 1188 | 1199 | 1208 | 1226 | ||
| 119 | Qwen3.5-27Bqwen/qwen3.5-27b | Qwen | 32.4 | $0.26 / $2.60 | 6 of 6 benchmarks | 1216 | 1227 | 1241 | 1242 | 1151 | 1091 |
| 120 | Claude Sonnet 4anthropic/claude-sonnet-4 | Anthropic | 31.7 | — | 4 of 6 benchmarks | 1189 | 1191 | 1202 | 1226 | ||
| 121 | Gemini 2.5 Flash Lite (thinking)google/gemini-2.5-flash-lite:thinking | 31.6 | $0.10 / $0.40 | 6 of 6 benchmarks | 1189 | 1190 | 1188 | 1194 | 1183 | 1183 | |
| 122 | Claude 3.7 Sonnetanthropic/claude-3-7-sonnet | Anthropic | 31.0 | — | 4 of 6 benchmarks | 1175 | 1187 | 1194 | 1230 | ||
| 123 | GPT-5.4 Nano (high)openai/gpt-5.4-nano:high | OpenAI | 30.5 | $0.20 / $1.25 | 5 of 6 benchmarks | 1199 | 1213 | 1226 | 1233 | 1117 | |
| 124 | Gemini 2.5 Flash Lite (nothinking)google/gemini-2.5-flash-lite:nothinking | 28.6 | $0.10 / $0.40 | 4 of 6 benchmarks | 1174 | 1185 | 1182 | 1198 | |||
| 125 | Hunyuan Vision 1.5 (thinking)tencent/hunyuan-vision-1.5:thinking | Tencent | 27.5 | — | 4 of 6 benchmarks | 1154 | 1158 | 1184 | 1227 | ||
| 126 | Claude 3.5 Sonnetanthropic/claude-3-5-sonnet | Anthropic | 27.1 | — | 4 of 6 benchmarks | 1161 | 1181 | 1186 | 1171 | ||
| 127 | GLM 4.6Vz-ai/glm-4.6v | Z.ai | 26.0 | $0.30 / $0.90 | 4 of 6 benchmarks | 1161 | 1172 | 1190 | 1155 | ||
| 128 | Gemini 2.0 Flash 001google/gemini-2.0-flash-001 | 26.0 | — | 5 of 6 benchmarks | 1172 | 1167 | 1177 | 1194 | 1150 | ||
| 129 | Hunyuan Large Visiontencent/hunyuan-large-vision | Tencent | 24.9 | — | 3 of 6 benchmarks | 1150 | 1146 | 1116 | |||
| 130 | GPT-5 Nano (high)openai/gpt-5-nano:high | OpenAI | 24.7 | $0.05 / $0.40 | 4 of 6 benchmarks | 1147 | 1159 | 1154 | 1188 | ||
| 131 | GLM 4.5Vz-ai/glm-4.5v | Z.ai | 23.8 | $0.60 / $1.80 | 4 of 6 benchmarks | 1153 | 1158 | 1166 | 1165 | ||
| 132 | Step 3stepfun/step-3 | StepFun | 23.1 | — | 4 of 6 benchmarks | 1144 | 1150 | 1145 | 1174 | ||
| 133 | Llama 4 Maverick 17B 128e Instructmeta-llama/llama-4-maverick-17b-128e-instruct | Meta | 22.8 | — | 4 of 6 benchmarks | 1146 | 1154 | 1152 | 1166 | ||
| 134 | Step 1o Turbo 202506stepfun/step-1o-turbo-202506 | StepFun | 22.8 | — | 4 of 6 benchmarks | 1155 | 1151 | 1162 | 1130 | ||
| 135 | Gemma 3 27Bgoogle/gemma-3-27b-it | 22.4 | $0.08 / $0.45 | 6 of 6 benchmarks | 1160 | 1162 | 1165 | 1169 | 1170 | 1119 | |
| 136 | Molmo 2 8Ballenai/molmo-2-8b | Allen Institute for AI | 22.3 | — | 3 of 6 benchmarks | 1113 | 1094 | 1118 | |||
| 137 | Mistral Medium 2508mistralai/mistral-medium-2508 | Mistral AI | 21.9 | — | 6 of 6 benchmarks | 1160 | 1173 | 1180 | 1174 | 1129 | 1128 |
| 138 | Llama 4 Scout 17B 16e Instructmeta-llama/llama-4-scout-17b-16e-instruct | Meta | 20.5 | — | 4 of 6 benchmarks | 1128 | 1131 | 1152 | 1151 | ||
| 139 | Claude 3.5 Haikuanthropic/claude-3-5-haiku | Anthropic | 20.0 | — | 4 of 6 benchmarks | 1127 | 1132 | 1131 | 1148 | ||
| 140 | Mistral Medium 2505mistralai/mistral-medium-2505 | Mistral AI | 19.2 | — | 6 of 6 benchmarks | 1155 | 1160 | 1167 | 1167 | 1125 | 1091 |
| 141 | Mistral Small 2506mistralai/mistral-small-2506 | Mistral AI | 17.9 | — | 6 of 6 benchmarks | 1140 | 1141 | 1140 | 1178 | 1134 | 1053 |
| 142 | Mistral Small 3.1 24B Instruct 2503mistralai/mistral-small-3.1-24b-instruct-2503 | Mistral AI | 17.8 | — | 6 of 6 benchmarks | 1127 | 1134 | 1129 | 1162 | 1130 | 1133 |
How this ranks
Every benchmark value becomes a percentile among the models that have it, so accuracy scores, Elo ratings and word error rates compare without hand-tuned scaling. Metrics where lower is better are inverted first. Raw values are never summed or averaged across benchmarks. A model's mean percentile is then shrunk toward the mean of the models that were broadly benchmarked, so a model tested twice cannot outrank a broadly tested one on two lucky results. Turning a data source off runs that same ranking code again in your browser over the sources you left on.
A model scored on fewer than 3 of the 6 ranked benchmarks in this category still ranks here, on the benchmarks it does have, and its row carries a partial coverage mark. On an equal score it sits under the model that earned the same number across more of the board.
Data sources
Turn a source off to drop every benchmark it feeds and re-rank the board in your browser. Your choice follows you across the leaderboard pages.
- LMArenaCC BY 4.0
Arena ratings by LMArena, from the public leaderboard dataset.