Consensus score
73.8
95% interval 65.4 to 80.2
73.8
95% interval 65.4 to 80.2
64th
IQR spread 47.9 pct
99
Across 8 domains
4
Newest observation 2d ago
Median within-board percentile by capability domain across boards with sufficient peer overlap.
Best verified run per board, with confidence intervals and Pareto frontier status.
| Benchmark | Rank | Score | Cost | Status | Observed |
|---|---|---|---|---|---|
| LMArena Vision LMArena · Chinese · high | #1 / 135Tied #1 | 1,411[1,379 to 1,443] | Unknown | Standard | 9 Nov 2026 |
| LMArena Vision (style control) LMArena · Chinese · high | #1 / 135Tied #1 | 1,379[1,347 to 1,412] | Unknown | Standard | 9 Nov 2026 |
| LMArena Vision (style control) LMArena · Diagram · high | #2 / 123Tied #1 | 1,342[1,327 to 1,356] | Unknown | Standard | 9 Nov 2026 |
| LMArena Vision LMArena · Diagram · high | #3 / 123Tied #1 | 1,353[1,338 to 1,367] | Unknown | Standard | 9 Nov 2026 |
| LMArena Vision LMArena · OCR · high | #3 / 123Tied #1 | 1,332[1,323 to 1,341] | Unknown | Standard | 9 Nov 2026 |
| LMArena Text (style control) LMArena · Russian · high | #4 / 385 | 1,516[1,508 to 1,524] | Unknown | Standard | 11 Nov 2026 |
| LMArena Vision LMArena · Homework · high | #4 / 118Tied #1 | 1,351[1,333 to 1,369] | Unknown | Standard | 9 Nov 2026 |
| LMArena Vision LMArena · Humor · high | #4 / 102Tied #1 | 1,329[1,294 to 1,364] | Unknown | Standard | 9 Nov 2026 |
| LMArena Vision (style control) LMArena · Homework · high | #4 / 118Tied #1 | 1,338[1,320 to 1,356] | Unknown | Standard | 9 Nov 2026 |
| LMArena Vision (style control) LMArena · Humor · high | #4 / 102Tied #1 | 1,305[1,269 to 1,340] | Unknown | Standard | 9 Nov 2026 |
| LMArena Vision (style control) LMArena · OCR · high | #4 / 123Tied #1 | 1,315[1,306 to 1,325] | Unknown | Standard | 9 Nov 2026 |
| HLE Epoch AI Benchmarking Hub · Default | #5 / 56 | 44.5%[40.7% to 48.4%] | Unknown | Standard | Not stated |
| LMArena Text LMArena · Math · high | #5 / 404Tied #1 | 1,525[1,514 to 1,537] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · Russian · high | #5 / 385 | 1,516[1,509 to 1,524] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Math · high | #5 / 404Tied #1 | 1,519[1,508 to 1,531] | Unknown | Standard | 11 Nov 2026 |
| GPQA diamond Epoch AI Benchmarking Hub · high | #6 / 338Tied #1 | 95.4%[92.6% to 98.1%] | Unknown | Standard | 2 Sept 2026 |
| LMArena Text LMArena · Industry life and physical and social science · high | #6 / 419 | 1,522[1,515 to 1,529] | Unknown | Standard | 11 Nov 2026 |
| LMArena Vision LMArena · Overall · high | #6 / 167Tied #1 | 1,317[1,309 to 1,325] | Unknown | Standard | 9 Nov 2026 |
| LMArena Text LMArena · Japanese · high | #7 / 288 | 1,515[1,493 to 1,536] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Japanese · high | #7 / 288 | 1,509[1,488 to 1,531] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · Chinese · high | #8 / 395 | 1,557[1,548 to 1,567] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · German · high | #8 / 320 | 1,503[1,483 to 1,523] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · Industry entertainment and sports and media · high | #8 / 419 | 1,484[1,478 to 1,490] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · Industry medicine and healthcare · high | #8 / 391 | 1,507[1,497 to 1,517] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · Non English · high | #8 / 421 | 1,492[1,487 to 1,497] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · Polish · high | #8 / 243Tied #1 | 1,509[1,489 to 1,529] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Polish · high | #8 / 243Tied #1 | 1,505[1,485 to 1,525] | Unknown | Standard | 11 Nov 2026 |
| LMArena Vision LMArena · English · high | #8 / 167Tied #1 | 1,315[1,302 to 1,329] | Unknown | Standard | 9 Nov 2026 |
| LMArena Vision (style control) LMArena · English · high | #8 / 167Tied #1 | 1,290[1,277 to 1,303] | Unknown | Standard | 9 Nov 2026 |
| EMB Vals AI benchmarks · high | #9 / 38Tied #1 | 72.9%[68.1% to 77.6%] | $8.24 | Standard | 15 Nov 2026 |
| Finance Agent Vals AI benchmarks · high | #9 / 38Tied #1 | 65.8%[61.8% to 69.8%] | $1.99 | Standard | 15 Nov 2026 |
| LMArena Document LMArena · Overall · high | #9 / 46 | 1,482[1,474 to 1,490] | Unknown | Standard | 9 Nov 2026 |
| LMArena Text LMArena · Creative writing · high | #9 / 419 | 1,493[1,487 to 1,500] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · Industry mathematical · high | #9 / 402 | 1,519[1,508 to 1,529] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · Multi turn · high | #9 / 419 | 1,503[1,496 to 1,510] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Industry writing and literature and language · high | #9 / 420 | 1,492[1,486 to 1,498] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Longer query · high | #9 / 399 | 1,511[1,505 to 1,516] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Overall · high | #9 / 421 | 1,497[1,492 to 1,501] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · Exclude ties · high | #10 / 421 | 1,515[1,510 to 1,521] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · Industry writing and literature and language · high | #10 / 420 | 1,496[1,490 to 1,502] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · Korean · high | #10 / 289Tied #1 | 1,470[1,449 to 1,491] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · Overall · high | #10 / 421 | 1,500[1,495 to 1,504] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Creative writing · high | #10 / 419 | 1,488[1,482 to 1,495] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Exclude ties · high | #10 / 421 | 1,514[1,508 to 1,520] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Industry entertainment and sports and media · high | #10 / 419 | 1,481[1,474 to 1,487] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Multi turn · high | #10 / 419 | 1,505[1,497 to 1,512] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Non English · high | #10 / 421 | 1,487[1,482 to 1,492] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · English · high | #11 / 421 | 1,500[1,495 to 1,506] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · Longer query · high | #11 / 399 | 1,508[1,503 to 1,514] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Hard prompts · high | #11 / 421 | 1,518[1,513 to 1,523] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Industry life and physical and social science · high | #11 / 419 | 1,518[1,511 to 1,525] | Unknown | Standard | 11 Nov 2026 |
| LMArena Vision LMArena · Creative writing vision · high | #11 / 106Tied #1 | 1,320[1,296 to 1,344] | Unknown | Standard | 9 Nov 2026 |
| LMArena Vision (style control) LMArena · Overall · high | #11 / 167 | 1,293[1,285 to 1,301] | Unknown | Standard | 9 Nov 2026 |
| LMArena Text LMArena · Expert · high | #12 / 371 | 1,527[1,519 to 1,535] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · Hard prompts · high | #12 / 421 | 1,510[1,505 to 1,515] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Chinese · high | #12 / 395 | 1,543[1,534 to 1,553] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · English · high | #12 / 421 | 1,500[1,494 to 1,505] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Hard prompts English · high | #12 / 419 | 1,517[1,511 to 1,523] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Industry mathematical · high | #12 / 402 | 1,512[1,501 to 1,523] | Unknown | Standard | 11 Nov 2026 |
| LMArena Vision (style control) LMArena · Creative writing vision · high | #12 / 106 | 1,294[1,270 to 1,318] | Unknown | Standard | 9 Nov 2026 |
| LMArena Text LMArena · Industry legal and government · high | #13 / 393 | 1,502[1,493 to 1,511] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · Instruction following · high | #13 / 421 | 1,493[1,487 to 1,499] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · Hard prompts English · high | #14 / 419 | 1,508[1,501 to 1,514] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · Industry business and management and financial operations · high | #14 / 414 | 1,483[1,476 to 1,490] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · Coding · high | #15 / 416 | 1,510[1,504 to 1,516] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · Industry software and IT services · high | #15 / 421 | 1,509[1,503 to 1,515] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · Spanish · high | #15 / 304 | 1,486[1,472 to 1,500] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Expert · high | #15 / 371 | 1,527[1,519 to 1,536] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Industry medicine and healthcare · high | #15 / 391Tied #1 | 1,508[1,498 to 1,518] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Instruction following · high | #15 / 421 | 1,491[1,486 to 1,497] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · German · high | #17 / 320 | 1,497[1,477 to 1,517] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text LMArena · French · high | #18 / 301 | 1,497[1,482 to 1,511] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Korean · high | #18 / 289Tied #1 | 1,458[1,437 to 1,480] | Unknown | Standard | 11 Nov 2026 |
| LMArena Document (style control) LMArena · Overall · high | #19 / 46 | 1,469[1,461 to 1,477] | Unknown | Standard | 9 Nov 2026 |
| LMArena WebDev LMArena · Image to webdev · high | #19 / 64 | 1,595[1,584 to 1,606] | Unknown | Standard | 15 Nov 2026 |
| Legal Agent Benchmark (HLAB-Vals) Vals AI benchmarks · high | #19 / 38 | 12.1%[7% to 17.1%] | $3.66 | Standard | 15 Nov 2026 |
| LMArena Text (style control) LMArena · Spanish · high | #20 / 304Tied #1 | 1,479[1,465 to 1,493] | Unknown | Standard | 11 Nov 2026 |
| Tax Agent Vals AI benchmarks · high | #20 / 38 | 32.1%[25.5% to 38.7%] | $0.86 | Standard | 15 Nov 2026 |
| LMArena Text (style control) LMArena · Coding · high | #21 / 416 | 1,531[1,525 to 1,537] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Industry business and management and financial operations · high | #23 / 414 | 1,485[1,478 to 1,492] | Unknown | Standard | 11 Nov 2026 |
| LMArena Text (style control) LMArena · Industry software and IT services · high | #23 / 421 | 1,521[1,515 to 1,526] | Unknown | Standard | 11 Nov 2026 |
| Vals Index Vals AI benchmarks · high | #24 / 38 | 56%[53.7% to 58.2%] | $5.03 | Standard | 15 Nov 2026 |
| ARC-AGI-2 Epoch AI Benchmarking Hub · high | #25 / 299 | 89.2% | Unknown | Standard | Not stated |
| Legal Research Vals AI benchmarks · high | #25 / 38 | 38.9%[32.3% to 45.6%] | $1.63 | Standard | 15 Nov 2026 |
| FrontierCode Epoch AI Benchmarking Hub · medium · devin-cli | #26 / 51 | 41.2% | Unknown | Standard | Not stated |
| LMArena Agent LMArena · Overall · high | #26 / 64 | 0[0 to 0] | $0.76 | Standard | 15 Nov 2026 |
| LMArena Text (style control) LMArena · French · high | #26 / 301 | 1,500[1,485 to 1,515] | Unknown | Standard | 11 Nov 2026 |
| Vibe Code Bench Vals AI benchmarks · high · OpenHands | #26 / 38 | 78.7%[71% to 86.3%] | $6.87 | Standard | 15 Nov 2026 |
| Terminal-Bench 4 Vals AI benchmarks · high | #27 / 38 | 19.7%[15.2% to 24.2%] | $8.65 | Standard | 15 Nov 2026 |
| LMArena Text (style control) LMArena · Industry legal and government · high | #28 / 393 | 1,494[1,485 to 1,504] | Unknown | Standard | 11 Nov 2026 |
| Code Migration Vals AI benchmarks · high | #31 / 38 | 38.2%[29.8% to 46.6%] | $8.31 | Standard | 15 Nov 2026 |
| LMArena WebDev LMArena · Webdev HTML · high | #34 / 155 | 1,608[1,594 to 1,623] | Unknown | Standard | 15 Nov 2026 |
| Vending-Bench 2 Epoch AI Benchmarking Hub · Default | #34 / 72 | $5,093.79 | Unknown | Standard | Not stated |
| LMArena WebDev LMArena · Overall · high | #42 / 156 | 1,583[1,576 to 1,590] | Unknown | Standard | 15 Nov 2026 |
| LMArena WebDev LMArena · Webdev · high | #42 / 156 | 1,583[1,576 to 1,590] | Unknown | Standard | 15 Nov 2026 |
| LMArena WebDev LMArena · Webdev react · high | #46 / 140 | 1,575[1,567 to 1,582] | Unknown | Standard | 15 Nov 2026 |
| ARC-AGI-2 Epoch AI Benchmarking Hub · medium | #51 / 299 | 82.9% | Unknown | Standard | Not stated |
| ARC-AGI-2 Epoch AI Benchmarking Hub · low | #57 / 299 | 77.5% | Unknown | Standard | Not stated |
| Bug Hunt Bench Bug Hunt Bench · high · Antigravity CLI | #97 / 145 | 18 | $11.03 | Standard | 15 Sept 2026 |
Same benchmark and harness measured across multiple reasoning effort settings.
Epoch AI Benchmarking Hub
77.5
82.9
89.2
Direct pairwise win rates for Gemini 3.8 Flash across shared benchmark boards.
23 shared boards
39% win
Head-to-head21 shared boards
29% win
Head-to-head20 shared boards
48% win
Head-to-head20 shared boards
40% win
Head-to-head20 shared boards
10% win
Head-to-head19 shared boards
79% win
Head-to-head19 shared boards
79% win
Head-to-head19 shared boards
58% win
Head-to-head18 shared boards
0% win
Head-to-head18 shared boards
56% win
Head-to-head