AnthropicUnited States· released Feb 2026

Claude Opus 4.6

Name: Claude Opus 4.6
Author: Anthropic

Best published result on 3 of the 22 tests it has taken.

Anthropic, United States. One of 12 models from this lab on the Index.

Third-party publishedHigh confidence (12+ comparable tests, all 4 Index categories)

Available in Theo as Theo Think

Theo orchestrates it for the steps it does best and always shows which engine answered.

Use it in Theo

Snapshot September 8, 2026

OpenCharts Index

75.4

Ranked #11 (10 to 16) of 36 · 24.6% behind the frontier on average · 22 comparable tests · 4 counting categories

Drop any single test and the Index lands between 72.6 and 78.4, anywhere from #10 to #16. Neighbours inside that range are ties.

7 of 8 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 Opus 4.6, in 22 results.

Strongest in agentic (91.1) and reasoning (83.5). It holds the best published result on Cybench, DeepResearch Bench and METR Time Horizon. It trails the frontier most in math, 39.4% behind on average. No comparable result yet in long context. Between 20% and 25% behind the best published results, on average. Drop any single test and it would sit anywhere from #10 to #16.

Best published result onCybench93.0%DeepResearch Bench55.3%METR Time Horizon12.0 h

Strongest category

Agentic

Long-horizon tasks with tools: terminals, desktops, browsers, whole jobs.

91.1

Cybench93.0% · best
DeepResearch Bench55.3% · best
METR Time Horizon12.0 h · 95% of best
Terminal-Bench79.8% · 94% of best
Vending-Bench 2$8,018 · 89% of best
APEX-Agents32.4% · 68% of best

Mean of 6 of the category's 8 comparable tests. Counts toward the Index.

Best single result

Cybench

Solve professional capture-the-flag security challenges without guidance.

93.0%

Best published result of 7 models.

Offensive security work demands systematic investigation and tool use under uncertainty.

Where it trails

FrontierMath Tier 4

Solve the hardest FrontierMath tier: problems that take expert mathematicians days.

26.8%

72.5% behind the best published result (97.6%) · #25 of 51.

The deepest end of the set. Progress here signals genuinely new capability.

Not measured yet

No comparable result yet in this category. A blank is a blank, never a zero, and it does not lower the Index.

Long context

Covered, but not averaged into the Index:

Knowledgesingle-test category, shown beside the Index · 62.2
Multimodalsingle-test category, shown beside the Index · 95.9
Human preferencesingle-test category, shown beside the Index · 99.3

Every score, with its source

See it at work

What Claude Opus 4.6 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.

Human preference · mean 99.3 across 1 comparable test · a single-test category, shown beside the Index and never averaged into it

Text ArenaLeaderboard

Answer the same prompt as a rival model; real users vote blind on which answer they prefer.

1,505

50% expected win rate against the board leader (1,507)
#2 of 92

high · board 2026-09-02

Source: Arena

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 Opus 4.6 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 Opus 4.6

#11·Index 75.4B

0 : 12

wins · 0 ties · 12 shared

GPT-6 Astra

#1·Index 99.0A+

GPT-6 Astra comes out ahead on 12 of the 12 tests both have taken.

Category means

Reasoning
83.5
99.6
Knowledge
62.2
100.0
Coding
66.6
97.7
Math
60.6
99.8
Agentic
91.1
98.5
Multimodal
95.9
—
Human preference
99.3
—

Shared tests · biggest gaps first

FrontierMath Tier 4Math
26.8%
97.6%
−72.5
Code Arena (WebDev)Coding
1,546
1,797
−61.8
FrontierCodeCoding
26.6%
53.3%
−49.8
ProofBenchMath
50.0%
99.0%
−49.0
SimpleQA VerifiedKnowledge
47.0%
75.6%
−37.8
APEX-AgentsAgentic
32.4%
46.7%
−30.1
FrontierMath (Tiers 1–3)Math
66.0%
93.7%
−29.6
ARC-AGI-2Reasoning
69.2%
95.0%
−27.2
WeirdMLCoding
78.0%
92.9%
−16.1
GPQA DiamondReasoning
90.5%
95.8%
−7.4
Mock AIME 2024–2025Math
94.4%
100.0%
−5.6
ARC-AGI-1Reasoning
94.0%
98.5%
−4.6

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 Astra

Scores

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.

Category means

Reasoning83.5 · 5/6
Knowledgebeside62.2 · 1/1
Coding66.6 · 4/7
Math60.6 · 4/5
Agentic91.1 · 6/8
Multimodalbeside95.9 · 1/1
Human preferencebeside99.3 · 1/1
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

90.5%

92.6% of the best (95.8%), above chance
65th percentile of 78 · n = 198

32k

Claude Opus 4.6

Claude Opus 4.6, in 22 results.

What Claude Opus 4.6 was asked to do, and how it did.

Claude Opus 4.6 against whoever you pick.

Every test, every source

Compare with its neighbours.

The best models, ranked here, working inside Theo.