Task ranking
Best LLM for Math & Reasoning
Claude Fable 5.1 ranks first for Math & Reasoning as of October 20, 2026.[1][2][3][4] It scores 100% on the OTIS Mock AIME 2024-2025 set in Epoch AI's independent run (#1 of 113), competition mathematics. Claude Opus 5.5 scores within two points, so read the top two as close rather than as a measured gap. Claude Opus 5.5 is the next-ranked alternative. It leads LiveBench Mathematics at 97.08% (#1 of 59), ranks #1 on LMArena instruction-following (Elo 1536), and solves 91.2% of FrontierMath tiers 1 to 3 for research-level mathematics.
About this recommendation
- Updated
- Evidence through
- Sources
- 25
- Revision
- v107
Decision audit
Why this result
Inspect the inputs and the computed order behind the recommendation.
Models screened
24
live candidates
Evaluation feeds
7
task-weighted
Winner coverage
88%
intended feed weight
Largest provider share
3 of 8
Anthropic
Anthropic: Claude Fable 5.1 leads the measured task comparison with 3 independent evaluation families. Community volume, release age and usage popularity add no ranking points. The leading scores are close under this policy.
Real-world track record is still developing. The recommendation can lead on independent measurements while practitioner reports accumulate.
Sources evaluated
These are the weights used in scoring. Missing measurements are omitted from each model’s average. A feed with no result for any model released in the last 60 days is stale and does not score until it catches up.
| Evaluation feed | Weight | Winner result | Field measured |
|---|---|---|---|
| LiveBench Mathematics (2026-06-25)Observed 2026-10-18 · evaluation date not supplied · claude-fable-5-1-max-effortfailing; last successful check 2026-10-18 | 28% | #2 | 18/24 |
| LiveBench Reasoning (2026-06-25)Observed 2026-10-18 · evaluation date not supplied · claude-fable-5-1-max-effortfailing; last successful check 2026-10-18 | 20% | #4 | 18/24 |
| LMArena MathObserved 2026-10-15 · evaluation date not supplied · max effort | 20% | #3 | 23/24 |
| GPQA Diamond (Epoch AI) | 12% | not measured | 21/24 |
| FrontierMath Tiers 1-3 (Epoch AI)Observed 2026-09-01 · evaluated 2026-09-01 · max effort · standard error 1.80 | 8% | #4 | 20/24 |
| LMArena Instruction FollowingObserved 2026-10-15 · evaluation date not supplied · max effort | 8% | #5 | 23/24 |
| OTIS Mock AIME (Epoch AI)Observed 2026-09-01 · evaluated 2026-09-01 · max effort · standard error 0.00 | 4% | #1 | 23/24 |
Provider concentration
Each exact model is scored separately; provider identity is not a ranking input.
- Anthropic3 models
- OpenAI3 models
- Google1 model
- Meta1 model
Decision table
Every published model is shown in computed order. Practitioner sources are distinct community threads, not the citations repeated in the prose below.
Anthropic · coverage 88%
#1 OTIS Mock AIME (Epoch AI) · #2 LiveBench Mathematics
Anthropic · coverage 100% · close score
#1 OTIS Mock AIME (Epoch AI) · #1 LiveBench Mathematics
Google · coverage 52% · close score
#1 GPQA Diamond (Epoch AI) · #1 OTIS Mock AIME (Epoch AI)
OpenAI · coverage 100%
#1 FrontierMath Tiers 1-3 (Epoch AI) · #1 OTIS Mock AIME (Epoch AI)
Anthropic · coverage 100%
#1 OTIS Mock AIME (Epoch AI) · #3 GPQA Diamond (Epoch AI)
OpenAI · coverage 100%
#1 FrontierMath Tiers 1-3 (Epoch AI) · #1 OTIS Mock AIME (Epoch AI)
Meta · coverage 88%
#9 LiveBench Mathematics · #12 OTIS Mock AIME (Epoch AI)
OpenAI · coverage 100%
#9 LiveBench Reasoning · #10 FrontierMath Tiers 1-3 (Epoch AI)
Task score averages the available, task-weighted measurements. Benchmark positions compare models matched to our catalog. Models measured by at least two independent evaluator families are considered first. Missing tests, practitioner report counts, popularity and release age add no points. Coverage describes evidence breadth, not a probability of correctness. Scores within two points are shown as close under this policy; this is not a statistical confidence interval. When the leading scores are that close, the model measured by more evaluator families ranks first.
- 01
Claude Fable 5.1Anthropic · $10 in / $50 out per M tokens · 1M context
Claude Fable 5.1 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. - 02
Claude Opus 5.5Anthropic · $4 in / $20 out per M tokens · 1M context
Claude Opus 5.5 leads LiveBench Mathematics at 97.08% (#1 of 59), ranks #1 on LMArena instruction-following (Elo 1536), and solves 91.2% of FrontierMath tiers 1 to 3 for research-level mathematics.
Best when: Choose for the highest LiveBench Mathematics performance at 97.08% (#1 of 59) alongside perfect 100% on OTIS Mock AIME 2024-2025.More tips
- Use when instruction-following matters: it ranks #1 of 152 on LMArena's instruction-following category with Elo 1536.
- Deploy for graduate-level science questions where it scores 90.6% on GPQA Diamond.
No practitioner reports are linked yet, so no cautions are listed. Absence of reports is not evidence of reliability.
- 03
Gemini 4 ArgonGoogle · $2 in / $10 out per M tokens · 1.0M context
Gemini 4 Argon dominates LMArena rankings with #1 overall text arena (Elo 1524), #1 maths category (Elo 1536), and #1 instruction-following (Elo 1536).
Best when: Select for human preference on math prompts where it ranks #1 of 148 on LMArena's maths category with Elo 1536.More tips
- Use for graduate-level science questions where it leads GPQA Diamond at 96.4% (#1 of 113).
- Deploy for competition mathematics with perfect 100% on OTIS Mock AIME 2024-2025.
Watch out for
- Note lower research-level mathematics coverage: it solves 86% of FrontierMath tiers 1 to 3 (#10 of 73), trailing top performers in this specific benchmark.
- 04
GPT-6.1 SolOpenAI · $2 in / $10 out per M tokens · 1.1M context
GPT-6.1 Sol leads FrontierMath research-level mathematics at 93.7% (#1 of 73), scores 96.83% on LiveBench Mathematics (#3 of 59), and achieves 100% on OTIS Mock AIME 2024-2025.
Best when: Prioritize for research-level mathematics where it solves 93.7% of FrontierMath tiers 1 to 3, ranking #1 of 73.More tips
- Use for graduate-level science questions with 95.4% on GPQA Diamond (#4 of 113).
- Deploy for competition mathematics with 96.83% on LiveBench Mathematics (#3 of 59) and perfect 100% on OTIS Mock AIME 2024-2025.
Watch out for
- Expect weaker human preference rankings on math-specific prompts: it places #34 of 148 on LMArena's maths category with Elo 1470, far below top performers.
- 05
Claude Sonnet 5.5Anthropic · $2 in / $10 out per M tokens · 1M context
Claude Sonnet 5.5 scores 100% on OTIS Mock AIME 2024-2025, 95.6% on GPQA Diamond (#3 of 113), and 96.13% on LiveBench Mathematics (#7 of 59), offering strong graduate science and competition math performance.
Best when: Use for graduate-level science questions where it scores 95.6% on GPQA Diamond (#3 of 113).More tips
- Deploy for competition mathematics with perfect 100% on OTIS Mock AIME 2024-2025.
Watch out for
- Note it ranks #12 of 148 on LMArena's maths category with Elo 1499, below several alternatives for human-preferred math responses.
- 06
GPT-6 AstraOpenAI · $10 in / $50 out per M tokens · 1.1M context
GPT-6 Astra leads LiveBench Reasoning at 92.65% (#1 of 59), ties for #1 on FrontierMath research mathematics at 93.7%, and scores 95.8% on GPQA Diamond (#2 of 113) with perfect 100% on OTIS Mock AIME.
Best when: Select for reasoning tasks where it scores 92.65% on LiveBench Reasoning (#1 of 59).More tips
- Use for research-level mathematics with 93.7% on FrontierMath tiers 1 to 3 (#1 of 73, tied).
- Deploy for graduate-level science with 95.8% on GPQA Diamond (#2 of 113).
Watch out for
- Expect lower human preference on math prompts: it ranks #40 of 148 on LMArena's maths category with Elo 1465.
- 07
Muse Spark 1.3Meta · $1.25 in / $4.25 out per M tokens · 1.0M context
Muse Spark 1.3 scores 95.95% on LiveBench Mathematics (#9 of 59) and ranks #9 on LMArena overall text arena with Elo 1494, though it solves only 74.4% of FrontierMath tiers 1 to 3.
Best when: Use for competition mathematics with 95.95% on LiveBench Mathematics (#9 of 59).Watch out for
- Note limited research-level mathematics depth: it solves 74.4% of FrontierMath tiers 1 to 3 (#18 of 73), below top performers.
- 08
GPT-5.6 TerraOpenAI · $2 in / $12 out per M tokens · 1.1M context
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.No practitioner reports are linked yet, so no cautions are listed. Absence of reports is not evidence of reliability.
Frequently asked
- What is the best model for Math & Reasoning?
- Claude Fable 5.1 ranks first as of October 20, 2026. Claude Fable 5.1 scores 100% on the OTIS Mock AIME 2024-2025 set in Epoch AI's independent run (#1 of 113), competition mathematics.[1]
Sources
- 1
“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) · Benchmark · Sep 1, 2026 - 2
“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 · Benchmark · Oct 18, 2026 - 3
“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 · Benchmark · Oct 15, 2026 - 4
“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) · Benchmark · Sep 22, 2026 - 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) · Benchmark · Sep 22, 2026 - 6
“Within the 113 catalog models matched to this evaluation: Scores 90.6% on GPQA Diamond in Epoch AI's independent run (#30 of 113), graduate-level science questions.”
GPQA Diamond (Epoch AI) · Benchmark · Sep 22, 2026 - 7
“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 · Benchmark · Oct 15, 2026 - 8
“Within the 113 catalog models matched to this evaluation: Scores 96.4% on GPQA Diamond in Epoch AI's independent run (#1 of 113), graduate-level science questions.”
GPQA Diamond (Epoch AI) · Benchmark · Oct 13, 2026 - 9
“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) · Benchmark · Oct 13, 2026 - 10
“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) · Benchmark · Oct 13, 2026 - 11
“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) · Benchmark · Sep 29, 2026 - 12
“Within the 113 catalog models matched to this evaluation: Scores 95.4% on GPQA Diamond in Epoch AI's independent run (#4 of 113), graduate-level science questions.”
GPQA Diamond (Epoch AI) · Benchmark · Sep 29, 2026 - 13
“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 · Benchmark · Oct 18, 2026 - 14
“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) · Benchmark · Sep 29, 2026 - 15
“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 · Benchmark · Oct 15, 2026 - 16
“Within the 113 catalog models matched to this evaluation: Scores 95.6% on GPQA Diamond in Epoch AI's independent run (#3 of 113), graduate-level science questions.”
GPQA Diamond (Epoch AI) · Benchmark · Sep 29, 2026 - 17
“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) · Benchmark · Sep 29, 2026 - 18
“Within the 148 catalog models matched to this evaluation: Ranks #12 of 148 on LMArena's maths category (Elo 1499), based on blind human preference votes for maths prompts.”
LMArena maths category · Benchmark · Oct 15, 2026 - 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 · Benchmark · Oct 18, 2026 - 20
“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) · Benchmark · Aug 30, 2026 - 21
“Within the 113 catalog models matched to this evaluation: Scores 95.8% on GPQA Diamond in Epoch AI's independent run (#2 of 113), graduate-level science questions.”
GPQA Diamond (Epoch AI) · Benchmark · Aug 30, 2026 - 22
“Within the 148 catalog models matched to this evaluation: Ranks #40 of 148 on LMArena's maths category (Elo 1465), based on blind human preference votes for maths prompts.”
LMArena maths category · Benchmark · Oct 15, 2026 - 23
“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 · Benchmark · Oct 18, 2026 - 24
“Within the 73 catalog models matched to this evaluation: Solves 74.4% of FrontierMath tiers 1 to 3 in Epoch AI's independent run (#18 of 73), research-level mathematics.”
FrontierMath Tiers 1-3 (Epoch AI) · Benchmark · Sep 16, 2026 - 25
“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 · Benchmark · Oct 18, 2026
Rankings synthesized from community evidence and open benchmarks. See methodology. Not driven by vendor marketing.