SimpleQA VerifiedEpoch-run
Answer a short factual question exactly, without searching, or say you do not know.
94.0%
best published result
of 75 models
max
Source: Epoch AI (internal runs)AnthropicUnited States· released Nov 2026
Claude Fable 5.2
Best published result on 10 of the 14 tests it has taken.
Anthropic, United States. One of 16 models from this lab on the Index.
Third-party publishedMedium confidence (7+ comparable tests, 3+ Index categories)
Snapshot November 10, 2026
OpenCharts Index
92.7
Ranked #2 (1 to 3) of 80 · 7.3% behind the frontier on average · 14 comparable tests · 3 counting categories
Part of this Index is an estimate. On the 3 categories it is tested in, it averages 94.9. The rest is filled in from models tested in every category, because it has too few results to count there yet: Coding, Multimodal. Such estimates have been within 4.1 points about nine times in ten.
Its band runs from 88.6 to 96.8: where the Index could land once it is tested in every category, or if any single test were dropped. That covers #1 to #3, and neighbours inside that range are ties.
4 of 5 hardest tests · share of the frontier
Distance to today's frontier on the hardest tests. Not a measure of intelligence.
What it's known for
Claude Fable 5.2, in 14 results.
Strongest in math (99.9) and reasoning (99.0). It holds the best published result on ARC-AGI-1, Chess Puzzles and EBR-bench and 7 more. It trails the frontier most in agentic, 14.1% behind on average. No comparable results yet in coding and human preference. Between 5% and 10% behind the best published results, on average. Part of its Index is an estimate. It has too few results in coding and multimodal for those categories to count, so they are filled in from models tested in every category. On the categories that do count it averages 94.9. Across its band it could sit anywhere from #1 to #3.
Best published result onARC-AGI-198.5%Chess Puzzles79.0%EBR-bench90.5%FrontierMath (Tiers 1–3)95.1%FrontierMath Tier 4100.0%FrontierSWE v269.9%Furniture Assembly90.0%Mystery Game Puzzles90.0%ProofBench100.0%SimpleQA Verified94.0%
Strongest category
Math
Competition and research mathematics, graded on the final answer or the proof.
99.9
- FrontierMath (Tiers 1–3)95.1% · best
- FrontierMath Tier 4100.0% · best
- ProofBench100.0% · best
- Mock AIME 2024–202599.7% · 100% of best
Mean of 4 of the category's 4 comparable tests. Counts toward the Index.
Best single result
ARC-AGI-1
Solve the original Abstraction and Reasoning Corpus: small grids, a handful of examples, one hidden rule.
98.5%
Best published result of 67 models.
The first ARC set is close to saturated at the frontier, so it now shows whether smaller models can generalize at all.
Where it trails
SpireClimb
Win a run of a deck-building strategy game at its hardest setting, playing from screenshots with a mouse and a keyboard.
19.6%
42.3% behind the best published result (34.0%) · #4 of 9.
It tests using a computer the way a person does, and decisions whose cost only shows up much later.
Not measured yet
No comparable result yet in these 2 categories. A blank is a blank, never a zero, and it does not lower the Index.
- Coding
- Human preference
Covered, but not averaged into the Index:
- Knowledgesingle-test category, shown beside the Index · 100.0
- Multimodal1 of 4 tests · needs 2 · mean 100.0
See it at work
What Claude Fable 5.2 was asked to do, and how it did.
Every test it has taken, by category, with the kind of task it faced, who produced the number and where the result landed against the best published one. The examples are original and illustrative, never items from the datasets themselves.
Knowledge · mean 100.0 across 1 comparable test · a single-test category, shown beside the Index and never averaged into it
Every result is read against the best published one on its test: percent scores as a share of the best above chance, Arena ratings as the expected win rate against the board leader, open-ended values on a log scale. Sitting a harder exam never lowers a model. Every number links to the publisher that produced it.
Head to head
Claude Fable 5.2 against whoever you pick.
Choose a rival. Every counted test both have taken appears side by side as a share of the best published result on that test, with the category means above. A win is a higher share; a gap under 2 points is a tie, because one item on a 45-problem exam is 2.2 points and publishers' own standard errors are of that order.
Claude Fable 5.2
#2·Index 92.7A
7 : 2
wins · 5 ties · 14 shared
GPT-6 Astra
#1·Index 93.7A
Claude Fable 5.2 comes out ahead on 7 of the 14 tests both have taken.
Category means
Reasoning
99.096.1Knowledge
100.080.4Coding
—93.4Math
99.998.8Agentic
85.990.6Multimodal
100.089.6Human preference
—85.2
Shared tests · biggest gaps first
- SpireClimbAgentic−42.319.6%34.0%
- SimpleQA VerifiedKnowledge+19.694.0%75.6%
- Furniture AssemblyMultimodal+16.790.0%80.0%
- EBR-benchAgentic+15.890.5%76.2%
- Mystery Game PuzzlesReasoning+7.490.0%84.0%
- Chess PuzzlesReasoning+6.879.0%74.0%
- FrontierSWE v2Agentic+6.369.9%65.5%
- GPQA DiamondReasoning−2.693.9%95.8%
- +2.4100.0%97.6%
- tie95.1%93.7%
- ARC-AGI-2Reasoningtie93.8%95.0%
- ProofBenchMathtie100.0%99.0%
- tie99.7%100.0%
- ARC-AGI-1Reasoningtie98.5%98.5%
Bars are distances to the best published result on each test (percent scores above chance, Arena ratings as win rate against the leader, open-ended values on a log scale). Raw values as the publishers report them.
Open GPT-6 AstraScores
Every test, every source
The best published run per benchmark, the raw value as the publisher reports it, who produced the number, how close it comes to the best published result on that test, and where the model sits among every model scored on it. Percent scores are read above chance, Arena ratings as a win rate against the board leader, open-ended values on a log scale. A test with too few models to compare against is shown but not counted, and so is an archived test: one whose publisher no longer covers new models.
Category means
- Reasoning99.0 · 5/8
- Knowledgebeside100.0 · 1/1
- Coding—
- Math99.9 · 4/4
- Agentic85.9 · 3/6
- Multimodal100.0 · 1/4
- Human preference—
- Long context—
Mean · tests taken of the category's comparable tests. A faded category either rests on fewer than half of them or is a single-test category (marked “beside”), which is shown next to the Index and never averaged into it.
GPQA DiamondEpoch-runReasoning
93.9%
96.5% of the best (96.4%), above chance
85th percentile of 96 · n = 198
max
Source: Epoch AI — AI Benchmarking HubARC-AGI-2LeaderboardReasoning
93.8%
98.7% of the best (95.0%)
97th percentile of 67 · n = 120
xhigh
Source: Epoch AI — AI Benchmarking HubARC-AGI-1LeaderboardReasoning
98.5%
best published result
of 67 models
xhigh
Source: Epoch AI — AI Benchmarking HubChess PuzzlesEpoch-runReasoning
79.0%
best published result
of 76 models
max
Source: Epoch AI — AI Benchmarking HubMystery Game PuzzlesEpoch-runReasoning
90.0%
best published result
of 72 models
max
Source: Epoch AI — AI Benchmarking HubSimpleQA VerifiedEpoch-runKnowledge
94.0%
best published result
of 75 models
max
Source: Epoch AI — AI Benchmarking HubMock AIME 2024–2025Epoch-runMath
99.7%
99.7% of the best (100.0%)
87th percentile of 98 · n = 45
max
Source: Epoch AI — AI Benchmarking HubFrontierMath (Tiers 1–3)Epoch-runMath
95.1%
best published result
of 78 models
max
Source: Epoch AI — AI Benchmarking HubFrontierMath Tier 4Epoch-runMath
100.0%
best published result
of 68 models
max
Source: Epoch AI — AI Benchmarking HubProofBenchLeaderboardMath
100.0%
best published result
of 81 models
max
Source: Epoch AI — AI Benchmarking HubFrontierSWE v2LeaderboardAgentic
69.9%
best published result
of 22 models
harness: proximus
Source: Proximal (FrontierSWE leaderboard)EBR-benchEpoch-runAgentic
90.5%
best published result
of 30 models
max
Source: Epoch AI — AI Benchmarking HubSpireClimbEpoch-runAgentic
19.6%
57.7% of the best (34.0%)
63th percentile of 9
max
Source: Epoch AI — AI Benchmarking HubFurniture AssemblyEpoch-runMultimodal
90.0%
best published result
of 37 models
max
Source: Epoch AI — AI Benchmarking HubNearby on the leaderboard
Compare with its neighbours.
The best models, ranked here, working inside Theo.
29 of the models on this page run inside Theo today. Theo picks the right one for each step and always shows which engine answered.