Benchmarks say one thing, people using the models say another. We pull both — plus real market pricing and measured speed — and score 227 models on one scale. 118 of them carry a machine score and a human score, which is where the disagreements show up.
28 of these models have a hands-on review on this site.
What feeds the index
live market pricing
benchmark indices
human preference votes
118 models scored by machines and humans
Benchmark indices cross-checked against human preference votes. The bar under each score splits it into the five things we weigh — same total, different shape means a different model for a different job.
Benchmarks, human votes, speed and real pricing on the same screen. The highlighted side wins that line.
Claude Fable 5.1
anthropic
89.9
intelligence · wins 6 lines
Claude Opus 5
anthropic
86.9
intelligence · wins 7 lines
Our composite score
| 60.9 | Overall | 62.5 |
|---|---|---|
| 90 | Intelligence40% of the total | 87 |
| 1 | Price / perf20% of the total | 14 |
| 85 | Capabilities15% of the total | 85 |
| 39 | Speed10% of the total | 41 |
| 55 | Ecosystem15% of the total | 55 |
Measured intelligence
| 56.8 | Intelligence indexArtificial Analysis | 54.1 |
|---|---|---|
| 81.6 | Coding indexArtificial Analysis | 78.0 |
| 58.2 | Agentic indexArtificial Analysis | 56.4 |
| 1514 · 3k votes | Human preferenceLMArena Elo | 1505 · 35k votes |
Speed
| 67 t/s | Output throughputtokens/second | 58 t/s |
|---|---|---|
| 266.51 s | Time to first token | 73.62 s |
Price per million tokens
| $10 | Input | $5.0 |
|---|---|---|
| $50 | Output | $25 |
| $40 | Blended 3:1typical real-world mix | $20 |
Capabilities
| 1M | Context window | 1M |
|---|---|---|
| 128k | Max output | 128k |
| Yes | Reasoning | Yes |
| No | Open weights | No |
| text, image, file | Modalities | text, image, file |
| 1 | Providers serving it | 1 |
| — | Knowledge cutoff | — |
Filter it down to the models that fit your constraints, then open a row for the full measurement set.
227 of 227 models
| # | Model | intelligence & profile | Intel. | Elo | t/s | In / Out | Context | Actions |
|---|---|---|---|---|---|---|---|---|
| 1 | Claude Fable 5.1 anthropic | 89.9 | 56.8 | 1514 | 67 | $10 / $50 | 1M | |
| 2 | anthropic | 86.9 | 54.1 | 1505 | 58 | $5.0 / $25 | 1M | |
| 3 | Claude Fable 5.1 (batch) anthropic | 84.0 | 56.8 | — | — | $5.0 / $25 | 1M | |
| 4 | anthropic | 84.0 | 53.2 | 1494 | 70 | $10 / $50 | 1M | |
| 5 | Kimi K3OPEN moonshotai | 80.5 | 50.2 | 1476 | 40 | $3.0 / $15 | 1.0M | |
| 6 | Claude Opus 5 (batch) anthropic | 80.5 | 54.1 | — | — | $2.5 / $13 | 1M | |
| 7 | Gemini 3.8 Flash google | 80.2 | 47.1 | 1495 | — | $0.75 / $3.8 | 1.0M | |
| 8 | GLM 5.3OPEN z-ai | 79.9 | 48.6 | 1474 | 84 | $1.4 / $4.4 | 1.3M | |
| 9 | Qwen3.8 Max (0902) qwen | 79.0 | 46.9 | 1480 | 40 | $2.0 / $6.0 | 1M | |
| 10 | Muse Spark 1.3 meta | 78.9 | 53.0 | — | 190 | $1.3 / $4.3 | 1.0M | |
| 11 | GPT-6 Astra (batch) openai | 78.9 | 54.7 | — | — | $5.0 / $25 | 1.1M | |
| 12 | GPT-6 Astra openai | 78.9 | 54.7 | — | 62 | $10 / $50 | 1.1M | |
| 13 | Gemini 3.7 Flash google | 78.2 | 45.2 | 1491 | 309 | $0.75 / $3.8 | 1.0M | |
| 14 | Qwen3.6 Max Preview qwen | 78.0 | — | 1446 | — | $1.0 / $6.2 | 262k | |
| 15 | anthropic | 77.9 | 44.3 | 1490 | 52 | $5.0 / $25 | 1M | |
| 16 | GLM 5.3 FlashOPEN z-ai | 77.8 | 46.2 | 1471 | 47 | $0.07 / $0.25 | 1.3M | |
| 17 | Claude Fable 5 (batch) anthropic | 77.5 | 53.2 | — | — | $5.0 / $25 | 1M | |
| 18 | x-ai | 76.6 | 50.6 | — | 64 | $2.0 / $6.0 | 500k | |
| 19 | GPT-5.6 Sol (batch) openai | 76.3 | 51.3 | — | — | $1.0 / $5.0 | 1.1M | |
| 20 | openai | 76.3 | 51.3 | — | 85 | $2.0 / $10 | 1.1M | |
| 21 | openai | 75.6 | 42.8 | 1470 | 149 | $2.5 / $15 | 1.1M | |
| 22 | openai | 75.6 | 45.6 | 1471 | 85 | $5.0 / $30 | 1.1M | |
| 23 | anthropic | 75.5 | 46.4 | 1462 | 62 | $5.0 / $25 | 1M | |
| 24 | Kimi K3 (batch)OPEN moonshotai | 75.2 | 50.2 | — | — | $3.0 / $15 | 1.0M | |
| 25 | meta | 73.7 | 43.3 | 1479 | — | $1.3 / $4.3 | 1.0M | |
| 26 | x-ai | 73.7 | 45.5 | 1453 | 58 | $2.0 / $6.0 | 500k | |
| 27 | GLM 5.2OPEN z-ai | 73.5 | 42.5 | 1467 | 71 | $0.97 / $3.0 | 1.0M | |
| 28 | 73.5 | 41.9 | 1484 | 206 | $1.5 / $9.0 | 1.0M | ||
| 29 | MiMo-V2.5OPEN xiaomi | 73.0 | — | 1428 | — | $0.14 / $0.28 | 1.1M | |
| 30 | anthropic | 72.6 | 45.1 | 1443 | 75 | $2.0 / $10 | 1M | |
| 31 | Gemini 3.6 Flash google | 72.3 | 40.3 | 1476 | 215 | $0.75 / $3.8 | 1.0M | |
| 32 | DeepSeek V4 Pro 0423OPEN deepseek | 71.8 | 42.1 | 1451 | 66 | $1.0 / $2.1 | 1.0M | |
| 33 | DeepSeek V4 Pro 0813OPEN deepseek | 71.8 | 42.1 | 1451 | 66 | $1.1 / $3.3 | 1.0M | |
| 34 | Qwen3.8 2.4T A95B (batch)OPEN qwen | 71.4 | 46.7 | — | — | $2.0 / $6.0 | 1.0M | |
| 35 | Qwen3.8 2.4T A95BOPEN qwen | 71.4 | 46.7 | — | 40 | $2.0 / $6.0 | 1.0M | |
| 36 | GLM 5.3 Flash (batch)OPEN z-ai | 71.3 | 46.2 | — | — | $0.15 / $0.50 | 1.0M | |
| 37 | Qwen3.8 27BOPEN qwen | 70.5 | 41.4 | 1438 | 44 | $0.42 / $3.0 | 1M | |
| 38 | GPT-5.6 Terra (batch) openai | 70.4 | 46.8 | — | — | $1.0 / $6.0 | 1.1M | |
| 39 | openai | 70.4 | 46.8 | — | 110 | $2.0 / $12 | 1.1M | |
| 40 | Gemini 3.8 Flash (batch) google | 69.6 | 47.1 | — | — | $0.38 / $1.9 | 1.0M | |
| 41 | Muse Spark 1.2 meta | 69.3 | 46.8 | — | 260 | $1.3 / $4.3 | 1.0M | |
| 42 | Qwen3.7 Max qwen | 69.3 | 37.0 | 1474 | 202 | $1.5 / $4.4 | 1M | |
| 43 | deepseek | 68.7 | 40.8 | 1432 | 134 | $0.09 / $0.18 | 1.0M | |
| 44 | DeepSeek V4 Flash 0731OPEN deepseek | 68.7 | 40.8 | 1432 | 134 | $0.14 / $0.28 | 1.3M | |
| 45 | anthropic | 68.5 | 30.8 | 1503 | 37 | $5.0 / $25 | 1M | |
| 46 | Claude Sonnet 5 (batch) anthropic | 68.0 | 45.1 | — | — | $1.0 / $5.0 | 1M | |
| 47 | 68.0 | 36.7 | 1480 | 115 | $2.0 / $12 | 1.0M | ||
| 48 | Gemini 3.7 Flash (batch) google | 66.5 | 45.2 | — | — | $0.38 / $1.9 | 1.0M | |
| 49 | GPT-5.6 Luna (batch) openai | 66.2 | 43.4 | — | — | $0.10 / $0.60 | 1.1M | |
| 50 | openai | 66.2 | 43.4 | — | 133 | $0.20 / $1.2 | 1.1M | |
| 51 | anthropic | 65.6 | 29.0 | 1458 | 44 | $3.0 / $15 | 1M | |
| 52 | MiMo-V2.5-ProOPEN xiaomi | 65.5 | 32.6 | 1465 | 35 | $0.43 / $0.87 | 1.1M | |
| 53 | Kimi K2.6OPEN moonshotai | 65.5 | 35.8 | 1455 | 46 | $0.95 / $4.0 | 262k | |
| 54 | Qwen3.5-Flash qwen | 65.2 | — | 1398 | — | $0.07 / $0.26 | 1M | |
| 55 | Qwen3.7 Plus qwen | 65.0 | 30.8 | 1454 | 56 | $0.32 / $1.3 | 1M | |
| 56 | Hy3OPEN tencent | 64.8 | 33.2 | 1440 | 95 | $0.13 / $0.53 | 262k | |
| 57 | Hy3 previewOPEN tencent | 64.8 | 33.2 | 1440 | 95 | $0.18 / $0.60 | 262k | |
| 58 | GLM 5.1OPEN z-ai | 64.8 | 31.9 | 1463 | 52 | $0.97 / $3.0 | 205k | |
| 59 | DeepSeek V4 Pro 0813 (batch)OPEN deepseek | 64.4 | 42.1 | — | — | $1.3 / $4.0 | 1.0M | |
| 60 | MiniMax M3OPEN minimax | 64.0 | 35.7 | 1435 | 86 | $0.30 / $1.2 | 1.0M | |
| 61 | DeepSeek V4 Flash Vision ExpOPEN deepseek | 63.6 | 41.5 | — | 120 | $0.22 / $0.66 | 1.0M | |
| 62 | Gemini 3 Flash Preview google | 63.6 | 30.7 | 1467 | 205 | $0.50 / $3.0 | 1.0M | |
| 63 | DeepSeek V4 Flash 0731 (batch)OPEN deepseek | 63.3 | 40.8 | — | — | $0.14 / $0.28 | 1.0M | |
| 64 | Qwen3.6 Plus qwen | 62.9 | 31.5 | 1437 | 55 | $0.33 / $1.9 | 1M | |
| 65 | GLM 5OPEN z-ai | 62.1 | 32.4 | 1446 | 69 | $0.60 / $1.9 | 205k | |
| 66 | GPT-5.1 openai | 61.4 | 29.1 | 1442 | 86 | $1.3 / $10 | 400k | |
| 67 | InklingOPEN thinkingmachines | 61.1 | 32.2 | 1439 | 73 | $1.0 / $4.0 | 1.0M | |
| 68 | Kimi K2.5OPEN moonshotai | 60.7 | 27.6 | 1446 | — | $0.45 / $2.3 | 262k | |
| 69 | openai | 60.1 | 35.0 | 1417 | 75 | $1.8 / $14 | 400k | |
| 70 | GPT-5.2 Chat openai | 60.1 | 35.0 | 1417 | 75 | $1.8 / $14 | 128k | |
| 71 | Gemini 3.6 Flash (batch) google | 58.8 | 40.3 | — | — | $0.38 / $1.9 | 1.0M | |
| 72 | GLM 4.7OPEN z-ai | 58.5 | 26.2 | 1436 | 99 | $0.40 / $1.8 | 205k | |
| 73 | Inkling SmallOPEN thinkingmachines | 58.1 | 32.3 | 1414 | 112 | $0.45 / $1.2 | 1.0M | |
| 74 | MiniMax M2.7OPEN minimax | 57.6 | 30.1 | 1405 | 71 | $0.30 / $1.2 | 205k | |
| 75 | GLM 5V Turbo z-ai | 57.5 | 27.6 | 1437 | — | $1.2 / $4.0 | 203k | |
| 76 | Gemini 3.5 Flash Lite google | 57.3 | 27.6 | 1436 | 371 | $0.30 / $2.5 | 1.0M | |
| 77 | GPT-5.4 Mini openai | 57.2 | 31.9 | 1412 | 156 | $0.75 / $4.5 | 400k | |
| 78 | GLM 4.6OPEN z-ai | 57.0 | 21.7 | 1440 | 50 | $0.43 / $1.8 | 205k | |
| 79 | DeepSeek V3.2OPEN deepseek | 56.4 | 25.3 | 1425 | — | $0.27 / $0.40 | 164k | |
| 80 | DeepSeek V3.2 ExpOPEN deepseek | 56.4 | 25.3 | 1425 | — | $0.27 / $0.41 | 164k | |
| 81 | Nova 2 Lite amazon | 55.9 | — | 1363 | — | $0.30 / $2.5 | 1M | |
| 82 | Qwen3 Max Thinking qwen | 55.9 | 25.0 | 1439 | — | $0.78 / $3.9 | 262k | |
| 83 | Qwen3.5 397B A17BOPEN qwen | 55.8 | 26.1 | 1438 | 91 | $0.39 / $2.3 | 262k | |
| 84 | Gemini 2.5 Pro google | 55.7 | 19.0 | 1458 | 128 | $1.3 / $10 | 1.0M | |
| 85 | Gemini 2.5 Pro Preview 06-05 google | 55.7 | 19.0 | 1458 | 128 | $1.3 / $10 | 1.0M | |
| 86 | Qwen3.5-122B-A10BOPEN qwen | 55.5 | 24.8 | 1418 | 133 | $0.29 / $2.4 | 262k | |
| 87 | Solar Pro 4 upstage | 55.3 | 32.7 | 1378 | 70 | $0.03 / $0.12 | 524k | |
| 88 | DeepSeek V3.1 TerminusOPEN deepseek | 54.8 | 23.5 | 1420 | — | $0.27 / $1.0 | 164k | |
| 89 | Kimi K2.7 CodeOPEN moonshotai | 54.8 | 33.9 | — | 45 | $0.66 / $3.4 | 262k | |
| 90 | GPT-5.4 Nano openai | 54.4 | 30.9 | 1373 | 172 | $0.20 / $1.3 | 400k | |
| 91 | Gemma 4 26B A4B OPEN google | 53.9 | 19.2 | 1435 | — | $0.07 / $0.34 | 262k | |
| 92 | Gemma 4 31BOPEN google | 53.6 | 22.2 | 1442 | 35 | $0.09 / $0.34 | 262k | |
| 93 | openai | 53.6 | 27.1 | 1406 | 89 | $1.3 / $10 | 400k | |
| 94 | Qwen3.5-27BOPEN qwen | 53.2 | 26.9 | 1408 | 76 | $0.20 / $1.6 | 262k | |
| 95 | Nex-N2-ProOPEN nex-agi | 53.0 | 32.7 | — | 129 | $0.25 / $1.0 | 262k | |
| 96 | MiniMax M3 (batch)OPEN minimax | 52.8 | 35.7 | — | — | $0.30 / $1.2 | 524k | |
| 97 | MiniMax M3 (free)OPEN minimax | 52.8 | 35.7 | — | — | Free / Free | 1.0M | |
| 98 | GPT-5.3-Codex openai | 52.7 | 36.9 | — | 134 | $1.8 / $14 | 400k | |
| 99 | x-ai | 52.1 | 29.3 | 1398 | 141 | $1.3 / $2.5 | 1M | |
| 100 | Mistral Medium 3.5 mistralai | 52.0 | 22.7 | 1421 | 129 | $1.5 / $7.5 | 262k | |
| 101 | MiniMax M1 minimax | 50.7 | — | 1343 | — | $0.55 / $2.2 | 1M | |
| 102 | Qwen3 Max qwen | 50.6 | 17.6 | 1439 | 48 | $0.78 / $3.9 | 262k | |
| 103 | Qwen3.5-35B-A3BOPEN qwen | 49.9 | 22.6 | 1396 | 152 | $0.31 / $1.3 | 262k | |
| 104 | MiniMax M2.1OPEN minimax | 49.3 | 24.6 | 1391 | 94 | $0.30 / $1.2 | 205k | |
| 105 | Gemini 3.1 Pro Preview (batch) google | 49.0 | 36.7 | — | — | $1.0 / $6.0 | 1.0M | |
| 106 | Claude Opus 4.5 anthropic | 48.0 | 33.6 | — | 53 | $5.0 / $25 | 200k | |
| 107 | Kimi K2 0711OPEN moonshotai | 47.5 | 25.9 | 1371 | 98 | $0.57 / $2.3 | 131k | |
| 108 | Step 3.5 FlashOPEN stepfun | 47.3 | 19.5 | 1404 | 217 | $0.10 / $0.30 | 262k | |
| 109 | GPT-5.2-Codex openai | 47.1 | 33.0 | — | — | $1.8 / $14 | 400k | |
| 110 | Gemini 3.1 Flash Lite google | 47.0 | 18.7 | 1415 | 292 | $0.25 / $1.5 | 1.0M | |
| 111 | Gemini 3.1 Flash Lite Preview google | 47.0 | 18.7 | 1415 | 292 | $0.25 / $1.5 | 1.0M | |
| 112 | MiniMax M2.5OPEN minimax | 46.7 | 26.8 | 1359 | 96 | $0.27 / $1.1 | 205k | |
| 113 | KAT-Coder-Pro V2 kwaipilot | 46.4 | 25.6 | — | — | $0.30 / $1.2 | 262k | |
| 114 | R1 0528OPEN deepseek | 46.4 | 13.9 | 1427 | — | $0.50 / $2.1 | 164k | |
| 115 | GLM 4.5OPEN z-ai | 46.3 | 13.3 | 1430 | — | $0.60 / $2.2 | 131k | |
| 116 | Inkling (free)OPEN thinkingmachines | 46.2 | 32.2 | — | — | Free / Free | 1.0M | |
| 117 | Inkling (batch)OPEN thinkingmachines | 46.2 | 32.2 | — | — | $1.0 / $4.0 | 524k | |
| 118 | DeepSeek V3.1OPEN deepseek | 46.1 | 14.8 | 1420 | — | $0.55 / $1.6 | 164k | |
| 119 | Qwen3 VL 235B A22B InstructOPEN qwen | 45.9 | 14.4 | 1420 | 57 | $0.21 / $1.9 | 262k | |
| 120 | Qwen3 VL 235B A22B ThinkingOPEN qwen | 45.9 | 14.4 | 1420 | 57 | $0.40 / $4.0 | 131k | |
| 121 | Qwen3 235B A22B Instruct 2507OPEN qwen | 45.8 | 13.1 | 1420 | 66 | $0.09 / $0.55 | 262k | |
| 122 | Gemini 2.5 Flash google | 45.1 | 13.9 | 1417 | 223 | $0.30 / $2.5 | 1.0M | |
| 123 | GLM 5 Turbo z-ai | 44.3 | 31.0 | — | — | $1.2 / $4.0 | 203k | |
| 124 | GPT-5 Mini openai | 43.7 | 24.2 | 1373 | 102 | $0.25 / $2.0 | 400k | |
| 125 | Qwen3 Next 80B A3B ThinkingOPEN qwen | 43.5 | 10.7 | 1417 | 191 | $0.15 / $1.2 | 262k | |
| 126 | Qwen3 Next 80B A3B InstructOPEN qwen | 43.5 | 10.7 | 1417 | 191 | $0.10 / $1.1 | 262k | |
| 127 | x-ai | 42.9 | 30.0 | — | 104 | $1.3 / $2.5 | 2M | |
| 128 | Qwen3.6 27BOPEN qwen | 42.8 | 29.0 | — | 55 | $0.30 / $2.0 | 262k | |
| 129 | Qwen3 235B A22B Thinking 2507OPEN qwen | 42.6 | 7.6 | 1415 | 62 | $0.23 / $2.3 | 131k | |
| 130 | Kimi K2 0905OPEN moonshotai | 42.5 | 17.2 | 1379 | 40 | $0.60 / $2.5 | 262k | |
| 131 | o1 openai | 42.1 | 17.1 | 1353 | — | $15 / $60 | 200k | |
| 132 | LongCat 2.0OPEN meituan | 41.3 | 25.8 | — | 50 | $0.30 / $1.2 | 1.0M | |
| 133 | Ling 3.0 FlashOPEN inclusionai | 41.2 | 27.4 | — | 328 | $0.02 / $0.06 | 262k | |
| 134 | MiniMax M2OPEN minimax | 40.8 | 21.7 | 1342 | 98 | $0.26 / $1.0 | 205k | |
| 135 | Qwen3.6 35B A3BOPEN qwen | 40.5 | 26.2 | — | 129 | $0.10 / $0.90 | 262k | |
| 136 | R1OPEN deepseek | 40.5 | 13.9 | 1373 | — | $0.70 / $2.5 | 64k | |
| 137 | GPT-5.1-Codex-Max openai | 39.9 | 27.9 | — | — | $1.3 / $10 | 400k | |
| 138 | GPT-5.1-Codex openai | 39.9 | 27.9 | — | — | $1.3 / $10 | 400k | |
| 139 | gpt-oss-120bOPEN openai | 39.3 | 15.6 | 1366 | 162 | $0.04 / $0.17 | 131k | |
| 140 | Gemini 3.5 Flash Lite (batch) google | 39.2 | 27.6 | — | — | $0.15 / $1.3 | 1.0M | |
| 141 | GLM 4.7 FlashOPEN z-ai | 38.5 | 16.6 | 1352 | 84 | $0.06 / $0.40 | 203k | |
| 142 | GLM 4.5 AirOPEN z-ai | 38.2 | 10.5 | 1384 | 74 | $0.13 / $0.85 | 131k | |
| 143 | DeepSeek V3 0324OPEN deepseek | 37.9 | 9.2 | 1375 | — | $0.25 / $1.0 | 164k | |
| 144 | Mercury 2 inception | 37.7 | 14.9 | 1358 | 821 | $0.25 / $0.75 | 128k | |
| 145 | GLM 4.6VOPEN z-ai | 37.4 | 10.7 | 1376 | 63 | $0.30 / $0.90 | 131k | |
| 146 | o3 Pro openai | 36.7 | 25.7 | — | — | $20 / $80 | 200k | |
| 147 | Step 3.7 FlashOPEN stepfun | 36.5 | 22.9 | — | 91 | $0.20 / $1.1 | 262k | |
| 148 | Trinity Large ThinkingOPEN arcee-ai | 35.8 | 12.0 | 1342 | 322 | $0.25 / $0.80 | 262k | |
| 149 | Qwen3 Coder 480B A35BOPEN qwen | 35.6 | 11.9 | 1356 | 57 | $0.30 / $1.0 | 262k | |
| 150 | GPT-5.1-Codex-Mini openai | 34.3 | 24.0 | — | — | $0.25 / $2.0 | 400k | |
| 151 | o3 openai | 33.9 | 23.7 | — | 116 | $2.0 / $8.0 | 200k | |
| 152 | Qwen3 235B A22BOPEN qwen | 33.9 | 7.6 | 1366 | 62 | $0.46 / $1.8 | 131k | |
| 153 | o3 Mini High openai | 33.7 | 12.9 | 1336 | 212 | $1.1 / $4.4 | 200k | |
| 154 | o3 Mini openai | 33.7 | 12.9 | 1336 | 212 | $1.1 / $4.4 | 200k | |
| 155 | Qwen3 30B A3B Instruct 2507OPEN qwen | 33.3 | 3.6 | 1384 | 106 | $0.05 / $0.19 | 262k | |
| 156 | Llama 3.3 70B InstructOPEN meta-llama | 32.6 | — | 1274 | — | $0.10 / $0.32 | 131k | |
| 157 | Kimi K2 ThinkingOPEN moonshotai | 32.4 | 25.9 | — | 98 | $0.60 / $2.5 | 262k | |
| 158 | DeepSeek V3OPEN deepseek | 32.2 | 8.3 | 1332 | — | $0.32 / $0.89 | 164k | |
| 159 | GPT-5 Nano openai | 32.1 | 13.7 | 1320 | 146 | $0.05 / $0.40 | 400k | |
| 160 | Claude Haiku 4.5 (batch) anthropic | 32.0 | 22.5 | — | — | $0.50 / $2.5 | 200k | |
| 161 | anthropic | 32.0 | 22.5 | — | — | $1.0 / $5.0 | 200k | |
| 162 | Granite 4.2 8BOPEN ibm-granite | 30.9 | 13.9 | 1318 | 143 | $0.10 / $0.15 | 131k | |
| 163 | Qwen3 32BOPEN qwen | 30.5 | 5.7 | 1340 | 103 | $0.08 / $0.28 | 131k | |
| 164 | Gemma 3 27BOPEN google | 30.1 | 2.0 | 1358 | — | $0.08 / $0.45 | 131k | |
| 165 | Llama 3.1 70B InstructOPEN meta-llama | 29.1 | — | 1261 | — | $0.40 / $0.40 | 131k | |
| 166 | o4 Mini High openai | 27.3 | 19.1 | — | 137 | $1.1 / $4.4 | 200k | |
| 167 | o4 Mini openai | 27.3 | 19.1 | — | 137 | $1.1 / $4.4 | 200k | |
| 168 | Qwen3 Coder NextOPEN qwen | 27.0 | 14.3 | — | 103 | $0.12 / $0.80 | 262k | |
| 169 | GLM 4.5VOPEN z-ai | 26.6 | 3.5 | 1333 | 64 | $0.60 / $1.8 | 66k | |
| 170 | Qwen3 30B A3B Thinking 2507OPEN qwen | 25.9 | 3.6 | 1317 | 106 | $0.20 / $2.4 | 82k | |
| 171 | Gemma 3 12BOPEN google | 25.8 | 1.0 | 1334 | — | $0.05 / $0.15 | 131k | |
| 172 | GPT-4o (2024-05-13) openai | 25.7 | 3.0 | 1300 | 94 | $5.0 / $15 | 128k | |
| 173 | Nemotron 3.5 LightningOPEN nvidia | 25.5 | 16.4 | — | 296 | $0.08 / $0.20 | 262k | |
| 174 | Qwen3.5-9BOPEN qwen | 24.7 | 14.8 | — | 82 | $0.10 / $0.15 | 262k | |
| 175 | Qwen3 30B A3BOPEN qwen | 24.5 | 3.6 | 1317 | 106 | $0.12 / $0.50 | 131k | |
| 176 | gpt-oss-20bOPEN openai | 24.4 | 9.1 | 1287 | 168 | $0.03 / $0.13 | 131k | |
| 177 | Mistral Small 4OPEN mistralai | 22.0 | 12.8 | — | 163 | $0.15 / $0.60 | 262k | |
| 178 | gpt-oss-120b (batch)OPEN openai | 21.8 | 15.6 | — | — | $0.15 / $0.60 | 131k | |
| 179 | Gemma 2 27BOPEN google | 21.4 | — | 1231 | — | $0.65 / $0.65 | 8k | |
| 180 | GPT-4o-mini (2024-07-18) openai | 20.7 | 1.4 | 1286 | 138 | $0.15 / $0.60 | 128k | |
| 181 | Command R+ (08-2024) cohere | 20.7 | — | 1229 | — | $2.5 / $10 | 128k | |
| 182 | GPT-4o (2024-08-06) openai | 20.2 | 3.9 | 1282 | 91 | $2.5 / $10 | 128k | |
| 183 | Gemma 3 4BOPEN google | 19.5 | 1.0 | 1291 | — | $0.05 / $0.10 | 131k | |
| 184 | GPT-4.1 openai | 18.9 | 13.2 | — | 161 | $2.0 / $8.0 | 1.0M | |
| 185 | o1-pro openai | 18.3 | 12.8 | — | — | $150 / $600 | 200k | |
| 186 | Devstral 2 2512OPEN mistralai | 17.9 | 6.6 | — | — | $0.40 / $2.0 | 262k | |
| 187 | Qwen3 VL 32B InstructOPEN qwen | 17.0 | 11.9 | — | 89 | $0.10 / $0.42 | 131k | |
| 188 | Sonar Reasoning Pro perplexity | 16.9 | 11.8 | — | — | $2.0 / $8.0 | 128k | |
| 189 | Mistral Large 2407 mistralai | 16.5 | 1.7 | 1266 | — | $2.0 / $6.0 | 131k | |
| 190 | GPT-4.1 Mini openai | 15.9 | 8.8 | — | 88 | $0.40 / $1.6 | 1.0M | |
| 191 | Mistral Medium 3.1 mistralai | 15.6 | 8.4 | — | 124 | $0.40 / $2.0 | 131k | |
| 192 | Mistral Small 3OPEN mistralai | 12.0 | 1.4 | 1233 | 141 | $0.05 / $0.08 | 33k | |
| 193 | Solar Pro 3 upstage | 11.1 | 8.5 | — | 130 | $0.15 / $0.60 | 131k | |
| 194 | Qwen3 Coder 30B A3B InstructOPEN qwen | 11.0 | 7.7 | — | 86 | $0.07 / $0.28 | 262k | |
| 195 | Llama 4 MaverickOPEN meta-llama | 10.8 | 8.5 | — | 82 | $0.20 / $0.70 | 1.0M | |
| 196 | Qwen3 VL 30B A3B InstructOPEN qwen | 10.7 | 7.5 | — | 110 | $0.15 / $0.60 | 262k | |
| 197 | Qwen3 VL 30B A3B ThinkingOPEN qwen | 10.7 | 7.5 | — | 110 | $0.20 / $2.4 | 262k | |
| 198 | Command AOPEN cohere | 10.5 | 2.1 | — | 57 | $2.5 / $10 | 256k | |
| 199 | GPT-4 Turbo openai | 10.2 | 2.3 | — | 33 | $10 / $30 | 128k | |
| 200 | GPT-4 Turbo Preview openai | 10.2 | 2.3 | — | 33 | $10 / $30 | 128k | |
| 201 | Command R (08-2024) cohere | 9.8 | — | 1187 | — | $0.15 / $0.60 | 128k | |
| 202 | Qwen3 14BOPEN qwen | 9.7 | 4.8 | — | 60 | $0.12 / $0.24 | 131k | |
| 203 | Llama 3.1 8B InstructOPEN meta-llama | 9.6 | — | 1186 | — | $0.05 / $0.08 | 131k | |
| 204 | Mistral Medium 3 mistralai | 9.6 | 6.7 | — | 131 | $0.40 / $2.0 | 131k | |
| 205 | Phi 4OPEN microsoft | 9.5 | 1.0 | 1217 | 37 | $0.07 / $0.14 | 16k | |
| 206 | Sonar perplexity | 8.4 | 5.9 | — | — | $1.0 / $1.0 | 127k | |
| 207 | Gemini 2.5 Flash Lite google | 8.0 | 5.6 | — | 373 | $0.10 / $0.40 | 1.0M | |
| 208 | GPT-4.1 Nano openai | 8.0 | 4.1 | — | 119 | $0.10 / $0.40 | 1.0M | |
| 209 | GPT-4o (2024-11-20) openai | 7.7 | 5.4 | — | 137 | $2.5 / $10 | 128k | |
| 210 | GPT-4o openai | 7.7 | 5.4 | — | 137 | $2.5 / $10 | 128k | |
| 211 | Llama 4 ScoutOPEN meta-llama | 7.4 | 4.6 | — | 87 | $0.10 / $0.30 | 1.3M | |
| 212 | Mistral Large 3 2512 mistralai | 7.2 | 1.0 | — | — | $0.50 / $1.5 | 262k | |
| 213 | Qwen3 VL 8B InstructOPEN qwen | 7.0 | 4.9 | — | 114 | $0.12 / $0.46 | 262k | |
| 214 | Qwen3 VL 8B ThinkingOPEN qwen | 7.0 | 4.9 | — | 114 | $0.18 / $2.1 | 131k | |
| 215 | GPT-4 openai | 6.3 | 1.5 | — | — | $30 / $60 | 8k | |
| 216 | Qwen3 8BOPEN qwen | 6.0 | 2.8 | — | 39 | $0.12 / $0.46 | 131k | |
| 217 | R1 Distill Llama 70BOPEN deepseek | 6.0 | 4.2 | — | 24 | $0.80 / $0.80 | 8k | |
| 218 | GPT-4o-mini openai | 5.6 | 1.4 | — | 138 | $0.15 / $0.60 | 128k | |
| 219 | Sonar Pro perplexity | 5.1 | 3.6 | — | — | $3.0 / $15 | 200k | |
| 220 | Granite 4.0 MicroOPEN ibm-granite | 1.4 | 1.0 | — | — | $0.02 / $0.11 | 131k | |
| 221 | Reka Flash 3OPEN rekaai | 1.4 | 1.0 | — | — | $0.10 / $0.20 | 66k | |
| 222 | Hermes 3 70B InstructOPEN nousresearch | 1.4 | 1.0 | — | 34 | $0.70 / $0.70 | 131k | |
| 223 | Saba mistralai | 1.4 | 1.0 | — | — | $0.20 / $0.60 | 33k | |
| 224 | Mistral Large mistralai | 1.4 | 1.0 | — | — | $2.0 / $6.0 | 128k | |
| 225 | Claude 3 Haiku anthropic | 1.4 | 1.0 | — | — | $0.25 / $1.3 | 200k | |
| 226 | Llama 3.2 3B InstructOPEN meta-llama | 0.0 | — | 1109 | — | $0.05 / $0.33 | 131k | |
| 227 | Llama 3.2 1B InstructOPEN meta-llama | 0.0 | — | 1054 | — | $0.03 / $0.20 | 60k |
The raw measurements come from third parties, credited below. What we add is the cross-source aggregation and the scoring: five axes on absolute 0-100 scales, weighted into one number. Absolute scales matter — a model scoring 78 today means the same as a model scoring 78 six months ago, so the history stays comparable as new models arrive.
Average of every published capability index we can cross-check — general intelligence, coding and agentic benchmarks.
Real market price per million tokens weighted against measured capability. A cheap weak model and an expensive strong one can score the same.
Context window, maximum output, modalities, reasoning support, prompt caching and how recent the knowledge cutoff is.
Measured output throughput (tokens per second) and time to first token. A reasoning model thinks before it answers, so its first token can take minutes — that is the trade it makes for capability, and it scores low here by design.
How many providers serve the model, and whether the weights are open — availability and independence from a single vendor.
Every collection is archived, so price cuts and ranking moves stay traceable. The sources publish today's numbers; this is the trail behind them.
The default ranking is capability, so the most expensive model can sit above a free one. Switch the criterion at the top of the page and the order changes completely — that is the point. Four questions come up constantly, and here are the answers.
Because the default ranking measures what a model can do, not what it costs. Price is a separate axis, and it has its own ranking one click away. Ranking on the composite score instead buried the frontier models behind cheap ones and turned the whole index into a price chart, which is not what a ranked list of models is for. Today that puts Anthropic: Claude Fable 5.1 first with 89.9 on capability at $50.00 per million output tokens.
Both count, equally. Benchmarks test verifiable problems; the arena records which answer people actually preferred. When a model carries both, each contributes half its intelligence score. The widest gap in the index right now is Qwen: Qwen3 235B A22B Thinking 2507: 11 from benchmarks against 70 from human votes — a 59-point disagreement that a single-source ranking would hide entirely.
The score itself, split into the five things we weigh. Each segment is that axis's contribution, so the full bar length is the composite score. Two models can share a total and have opposite shapes — one fast and cheap, the other slow and brilliant.
Because that measurement does not exist yet, and we would rather show a dash than invent a number. Of the 227 models listed, 214 carry benchmark indices and 131 carry human ratings. A missing value never counts as zero — it simply leaves that axis out of the model's score, and the “Double-validated” filter isolates the 118 models measured both ways.
A ranking of 227 large language models built by cross-referencing three independent sources: live market pricing from OpenRouter, benchmark indices from Artificial Analysis, and human preference votes from LMArena. We aggregate those measurements and compute our own score across five weighted axes. The raw measurements belong to their sources; the scoring and the ranking are ours.
By measured capability, not by price. A model that costs ten times more but answers harder questions ranks higher by default. Three alternative orderings are one click away — price/performance, speed, and the weighted composite score — because "best" means something different depending on whether you are paying the bill or waiting for the response.
Two things, weighted equally when both exist. Machine benchmarks contribute the general intelligence index (50%), the coding index (25%) and the agentic index (25%), each normalised against its own ceiling. Human preference contributes the LMArena Elo rating, mapped onto the same 0-100 scale. A model measured by only one of the two is scored on that one alone.
A model carrying both a machine benchmark score and a human preference rating. 118 of the 227 models in the index qualify. It matters because the two methods disagree more often than you would expect: benchmarks measure what is testable, the arena measures what people actually prefer after using the model.
They measure different things. Benchmarks test verifiable problems with known answers — maths, code that compiles, factual recall. Human votes reward tone, instruction-following, formatting and refusal behaviour, which no benchmark captures. A model can top the benchmarks and lose blind head-to-head votes, and the reverse happens just as often.
Because nearly every current model scored in the 90s on it, so it separated nothing. Worse, averaging it in rewarded models measured on maths and almost nothing else — one model with no coding or agentic score reached a higher intelligence score than models measured on everything. Artificial Analysis has since stopped publishing that index on the endpoint we use, which changes nothing for us: it was already out of the score.
Measured capability divided by the blended price per million tokens, weighted one part input to three parts output, which reflects typical real-world usage. The result is placed on a logarithmic scale because prices span four orders of magnitude. Free or unpriced models receive a neutral 50 rather than a perfect score, since they are not comparable.
Yes. They come from OpenRouter, which publishes live per-token rates charged by the providers actually serving each model, in US dollars per million tokens. Input and output are shown separately because the gap between them is often large and matters more than the headline figure.
Every 48 hours. Each collection re-fetches all three sources, recalculates every score, and writes a historical snapshot so price changes and ranking movements stay traceable. Models that disappear from the sources are marked inactive rather than deleted, so their history survives.
A model whose weights can be downloaded and self-hosted, as opposed to one reachable only through a vendor API. 115 of the 227 models qualify. Open weights add points on the ecosystem axis because they remove vendor lock-in, but a closed model served by many providers is not penalised to zero.
So the numbers stay comparable over time. If scores were normalised against the current field, every new model release would silently move every existing score, and a snapshot from six months ago would mean nothing. On an absolute scale, a model scoring 78 today means the same as a model scoring 78 last year.
28 of the 227 models link to a full review on this site, written after actually using the model. Those links appear directly on each model card and in the comparison table. The rest are scored from published measurements only, which is stated rather than hidden.
All raw metrics belong to their respective sources and are reproduced with attribution. The composite scores, weighting and rankings on this page are ThePlanetTools.ai's own computation and should be attributed to us, not to the data providers.
We're developers and SaaS builders who use these tools daily in production. Every review comes from hands-on experience building real products — DealPropFirm, ThePlanetIndicator, PropFirmsCodes, and many more. We don't just review tools — we build and ship with them every day.
Written and tested by developers who build with these tools daily.