Z.ai

7 sources, 6 independent of the vendor: centre 26; the sources disagree widely, so shown as a band.

• 34.4 AI Coding Daily LLM Coding Leaderboard: 54.82 · leaderboard
• Artificial Analysis · third party
• 26.7 Chess Puzzles (Epoch AI): 37 · third party
• 27.2 Epoch Capabilities Index (Epoch AI): 156.17 · third party
• 22.9 FrontierMath-Tier-4-v2-Private (Epoch AI): 29.27 · third party
• 30.1 FrontierMath-Tiers-1-3-v2-Private (Epoch AI): 68.77 · third party
• 18.8 GPQA Diamond: 88.13 · third party
• 27.4 Humanity's Last Exam: 62.5 · third party
• 26.2 Mystery Game Puzzles (Epoch AI): 33 · third party
• 25.4 OTIS Mock AIME 2024-2025 (Epoch AI): 91.11 · third party
• 22.5 SWE-bench Verified: 95.4 · third party
• 21.6 SimpleQA Verified (Epoch AI): 41 · third party
• 26.1 Terminal-Bench 2.1: 86.05 · vendor

how FTI is computed

5 sources, 4 independent of the vendor: centre 21.

• 16.6 Chess Puzzles (Epoch AI): 21 · third party
• 23.2 Epoch Capabilities Index (Epoch AI): 153.22 · third party
• 18 FrontierMath-Tier-4-v2-Private (Epoch AI): 17.07 · third party
• 22.1 FrontierMath-Tiers-1-3-v2-Private (Epoch AI): 55.79 · third party
• 17.8 GPQA Diamond: 86.36 · third party
• 24.3 Humanity's Last Exam: 55.3 · third party
• 11.1 Mystery Game Puzzles (Epoch AI): 8 · third party
• 26.7 OTIS Mock AIME 2024-2025 (Epoch AI): 93.89 · third party
• 21.4 SWE-bench Verified: 92 · third party
• 25.3 Terminal-Bench 2.1: 84.28 · vendor

how FTI is computed

5 sources, 1 independent of the vendor: centre 7.

• 6.7 Humanity's Last Exam: 14.4 · vendor
• 6.5 LiveCodeBench: 64 · vendor
• 10.7 OTIS Mock AIME 2024-2025 (Epoch AI): 58.33 · third party
• 7 SWE-bench Verified: 59.2 · vendor
• 9.4 tau-bench (airline + retail + telecom): 79.5 · vendor
• aime 2025: 91.6 · vendor · not used: only 0 reference models
• browsecomp: 42.8 · vendor · not used: no reference models for this benchmark
• gpqa: 75.2 · vendor · not used: no reference models for this benchmark

how FTI is computed