# Best LLM for Competition Math > Updated Oct 20, 2026 ยท evidence through Oct 20, 2026. Synthesized from cited community evidence and open benchmarks, not vendor marketing. Canonical page: https://modelmatchbook.com/best/math-competition Claude Fable 5.1 ranks first for Competition Math as of October 20, 2026. It scores 97.01% on LiveBench Mathematics (#2 of 59) and 100% on OTIS Mock AIME 2024-2025, placing it among the top performers on objective competition math benchmarks. Claude Opus 5.5 scores within two points, so read the top two as close rather than as a measured gap. Watch out: Watch for research-level math gaps: it solves only 90.2% of FrontierMath tiers 1-3 (#4 of 73), trailing top models on research mathematics. Claude Opus 5.5 is the next-ranked alternative. It leads LiveBench Mathematics at 97.08% (#1 of 59) while matching perfect scores on OTIS Mock AIME, making it the strongest on objective competition benchmarks. [1,2,3,4,5] ## 1. Claude Fable 5.1 https://modelmatchbook.com/models/anthropic/claude-fable-5.1-20260831 Claude Fable 5.1 scores 97.01% on LiveBench Mathematics (#2 of 59) and 100% on OTIS Mock AIME 2024-2025, placing it among the top performers on objective competition math benchmarks. **Best when:** Use for competition math problems where you need 97.01% accuracy on LiveBench Mathematics tasks, ranking #2 of 59 models tested. [1] **More tips:** - Deploy for AIME-style competition problems where it achieved 100% on the OTIS Mock AIME 2024-2025 set, tied for #1 of 113 models. [2] - Consider when human preference rankings matter: it places #3 of 148 on LMArena's maths category with Elo 1528. [6] **Watch out for:** - Watch for research-level math gaps: it solves only 90.2% of FrontierMath tiers 1-3 (#4 of 73), trailing top models on research mathematics. [3] ## 2. Claude Opus 5.5 https://modelmatchbook.com/models/anthropic/claude-opus-5.5-20260921 Claude Opus 5.5 leads LiveBench Mathematics at 97.08% (#1 of 59) while matching perfect scores on OTIS Mock AIME, making it the strongest on objective competition benchmarks. **Best when:** Prioritize for competition math where LiveBench Mathematics scores matter: 97.08% places it #1 of 59 models. [4] **More tips:** - Use for AIME-level problems where it scored 100% on OTIS Mock AIME 2024-2025, tied for #1 of 113. [5] - Consider when instruction-following quality affects math workflows: it ranks #1 of 152 on LMArena's instruction-following category with Elo 1536. [7] **Watch out for:** - Note the human preference gap in math-specific rankings: it places only #8 of 148 on LMArena's maths category (Elo 1508), below several competitors. [8] - Watch for research math limitations: 91.2% on FrontierMath tiers 1-3 (#3 of 73) leaves room versus top research-level performers. [9] ## 3. Gemini 4 Argon https://modelmatchbook.com/models/google/gemini-4-argon-20261013 Gemini 4 Argon dominates human preference rankings (#1 on LMArena maths with Elo 1536) while matching perfect AIME scores, though it lags on research-level FrontierMath. **Best when:** Use when human preference for math output quality matters: it ranks #1 of 148 on LMArena's maths category with Elo 1536. [10] **More tips:** - Deploy for AIME competition problems where it scored 100% on OTIS Mock AIME 2024-2025, tied for #1 of 113. [11] - Consider for overall text quality needs alongside math: it ranks #1 of 152 on LMArena's overall text arena with Elo 1524. [12] **Watch out for:** - Watch for research mathematics gaps: 86% on FrontierMath tiers 1-3 (#10 of 73) trails top performers by significant margin. [13] ## 4. GPT-6.1 Sol https://modelmatchbook.com/models/openai/gpt-6.1-sol-20260929 GPT-6.1 Sol leads FrontierMath research mathematics at 93.7% (#1 of 73) while scoring 96.83% on LiveBench Mathematics (#3 of 59), though human preference rankings lag significantly. **Best when:** Prioritize for research-level mathematics: 93.7% on FrontierMath tiers 1-3 ranks #1 of 73 models. [14] **More tips:** - Use for competition math where LiveBench scores matter: 96.83% places #3 of 59 on LiveBench Mathematics. [15] - Deploy for AIME problems where it scored 100% on OTIS Mock AIME 2024-2025, tied for #1 of 113. [16] **Watch out for:** - Expect lower human preference rankings for math: #34 of 148 on LMArena's maths category (Elo 1470) suggests outputs may feel less natural or helpful to human raters. [17] - Note instruction-following limitations: #24 of 152 on LMArena's instruction-following category (Elo 1469) may affect complex multi-step math workflows. [18] ## 5. GPT-6 Astra https://modelmatchbook.com/models/openai/gpt-6-astra-20260903 GPT-6 Astra scores 92.65% on LiveBench Reasoning (#1 of 59), an objective, ground-truth-scored evaluation refreshed with new questions. **Best when:** Its strongest result for this task is on LiveBench Reasoning, which measures multi-step reasoning. [19] ## 6. Claude Sonnet 5.5 https://modelmatchbook.com/models/anthropic/claude-sonnet-5.5-20260928 Claude Sonnet 5.5 scores 100% on the OTIS Mock AIME 2024-2025 set in Epoch AI's independent run (#1 of 113), competition mathematics. **Best when:** Its strongest result for this task is on OTIS Mock AIME, which measures competition mathematics. [20] ## 7. GPT-5.6 Terra https://modelmatchbook.com/models/openai/gpt-5.6-terra-20260709 GPT-5.6 Terra scores 90.63% on LiveBench Reasoning (#9 of 59), an objective, ground-truth-scored evaluation refreshed with new questions. **Best when:** Its strongest result for this task is on LiveBench Reasoning, which measures multi-step reasoning. [21] ## 8. Muse Spark 1.3 https://modelmatchbook.com/models/meta/muse-spark-1.3-20260902 Muse Spark 1.3 scores 95.95% on LiveBench Mathematics (#9 of 59), using objectively scored competition and olympiad-style tasks. **Best when:** Its strongest result for this task is on LiveBench Mathematics, which measures mathematical problem solving. [22] ## FAQ **What is the best model for Competition Math?** Claude Fable 5.1 ranks first as of October 20, 2026. Claude Fable 5.1 scores 97.01% on LiveBench Mathematics (#2 of 59) and 100% on OTIS Mock AIME 2024-2025, placing it among the top performers on objective competition math benchmarks. [1,2] **What should I watch out for with Claude Fable 5.1?** Watch for research-level math gaps: it solves only 90.2% of FrontierMath tiers 1-3 (#4 of 73), trailing top models on research mathematics. [3] **What is the best alternative to Claude Fable 5.1?** Claude Opus 5.5 ranks second. Claude Opus 5.5 leads LiveBench Mathematics at 97.08% (#1 of 59) while matching perfect scores on OTIS Mock AIME, making it the strongest on objective competition benchmarks. [4,5] ## Sources 1. "Within the 59 catalog models matched to this evaluation: Scores 97.01% on LiveBench Mathematics (#2 of 59), using objectively scored competition and olympiad-style tasks." (LiveBench Mathematics, https://livebench.ai/) 2. "Within the 113 catalog models matched to this evaluation: Scores 100% on the OTIS Mock AIME 2024-2025 set in Epoch AI's independent run (#1 of 113), competition mathematics." (OTIS Mock AIME (Epoch AI), https://epoch.ai/benchmarks) 3. "Within the 73 catalog models matched to this evaluation: Solves 90.2% of FrontierMath tiers 1 to 3 in Epoch AI's independent run (#4 of 73), research-level mathematics." (FrontierMath Tiers 1-3 (Epoch AI), https://epoch.ai/benchmarks) 4. "Within the 59 catalog models matched to this evaluation: Scores 97.08% on LiveBench Mathematics (#1 of 59), using objectively scored competition and olympiad-style tasks." (LiveBench Mathematics, https://livebench.ai/) 5. "Within the 113 catalog models matched to this evaluation: Scores 100% on the OTIS Mock AIME 2024-2025 set in Epoch AI's independent run (#1 of 113), competition mathematics." (OTIS Mock AIME (Epoch AI), https://epoch.ai/benchmarks) 6. "Within the 148 catalog models matched to this evaluation: Ranks #3 of 148 on LMArena's maths category (Elo 1528), based on blind human preference votes for maths prompts." (LMArena maths category, https://lmarena.ai/leaderboard) 7. "Within the 152 catalog models matched to this evaluation: Ranks #1 of 152 on LMArena's instruction-following category (Elo 1536), based on blind human preference votes." (LMArena instruction-following category, https://lmarena.ai/leaderboard) 8. "Within the 148 catalog models matched to this evaluation: Ranks #8 of 148 on LMArena's maths category (Elo 1508), based on blind human preference votes for maths prompts." (LMArena maths category, https://lmarena.ai/leaderboard) 9. "Within the 73 catalog models matched to this evaluation: Solves 91.2% of FrontierMath tiers 1 to 3 in Epoch AI's independent run (#3 of 73), research-level mathematics." (FrontierMath Tiers 1-3 (Epoch AI), https://epoch.ai/benchmarks) 10. "Within the 148 catalog models matched to this evaluation: Ranks #1 of 148 on LMArena's maths category (Elo 1536), based on blind human preference votes for maths prompts." (LMArena maths category, https://lmarena.ai/leaderboard) 11. "Within the 113 catalog models matched to this evaluation: Scores 100% on the OTIS Mock AIME 2024-2025 set in Epoch AI's independent run (#1 of 113), competition mathematics." (OTIS Mock AIME (Epoch AI), https://epoch.ai/benchmarks) 12. "Within the 152 catalog models matched to this evaluation: Ranks #1 of 152 on LMArena's overall text arena (Elo 1524), based on blind human preference votes." (LMArena text arena, https://lmarena.ai/leaderboard) 13. "Within the 73 catalog models matched to this evaluation: Solves 86% of FrontierMath tiers 1 to 3 in Epoch AI's independent run (#10 of 73), research-level mathematics." (FrontierMath Tiers 1-3 (Epoch AI), https://epoch.ai/benchmarks) 14. "Within the 73 catalog models matched to this evaluation: Solves 93.7% of FrontierMath tiers 1 to 3 in Epoch AI's independent run (#1 of 73), research-level mathematics." (FrontierMath Tiers 1-3 (Epoch AI), https://epoch.ai/benchmarks) 15. "Within the 59 catalog models matched to this evaluation: Scores 96.83% on LiveBench Mathematics (#3 of 59), using objectively scored competition and olympiad-style tasks." (LiveBench Mathematics, https://livebench.ai/) 16. "Within the 113 catalog models matched to this evaluation: Scores 100% on the OTIS Mock AIME 2024-2025 set in Epoch AI's independent run (#1 of 113), competition mathematics." (OTIS Mock AIME (Epoch AI), https://epoch.ai/benchmarks) 17. "Within the 148 catalog models matched to this evaluation: Ranks #34 of 148 on LMArena's maths category (Elo 1470), based on blind human preference votes for maths prompts." (LMArena maths category, https://lmarena.ai/leaderboard) 18. "Within the 152 catalog models matched to this evaluation: Ranks #24 of 152 on LMArena's instruction-following category (Elo 1469), based on blind human preference votes." (LMArena instruction-following category, https://lmarena.ai/leaderboard) 19. "Within the 59 catalog models matched to this evaluation: Scores 92.65% on LiveBench Reasoning (#1 of 59), an objective, ground-truth-scored evaluation refreshed with new questions." (LiveBench Reasoning, https://livebench.ai/) 20. "Within the 113 catalog models matched to this evaluation: Scores 100% on the OTIS Mock AIME 2024-2025 set in Epoch AI's independent run (#1 of 113), competition mathematics." (OTIS Mock AIME (Epoch AI), https://epoch.ai/benchmarks) 21. "Within the 59 catalog models matched to this evaluation: Scores 90.63% on LiveBench Reasoning (#9 of 59), an objective, ground-truth-scored evaluation refreshed with new questions." (LiveBench Reasoning, https://livebench.ai/) 22. "Within the 59 catalog models matched to this evaluation: Scores 95.95% on LiveBench Mathematics (#9 of 59), using objectively scored competition and olympiad-style tasks." (LiveBench Mathematics, https://livebench.ai/)