RankingEffort benchmarks

Claude Opus 4.6

Compare models

79 measured results · 9 effort settings · 56 benchmark/harness combinations

Model evidence and sources →

Change model

Model

Current configurationSources

Compare with

Choose a second model and its effort to see both profiles and benchmark differences below.

Compare withModel and configuration

Capability profile

One configuration, seven capabilities

Comparison score · 0–100

AClaude Opus 4.6 · Max3 of 7 capabilities supported

Gaps are unknown, not zero. Hollow points are preliminary.

Agentic
A38.4Supported
Hard reasoning
A44.7Supported
Coding
A40.4Preliminary
Human pref
Unknown
Knowledge
A27.8Supported
Multimodal
A24.9Preliminary
Long context
A41.8Preliminary

Each profile belongs to the named configuration. Capability scores use a shared panel for all axes, distinct from the overall ranking panel; they are not accuracy percentages or direct head-to-head wins.

Capability score evidence

Claude Opus 4.6 · Max: 40.8 overall comparison score. Meets the overall evidence requirements.

6 areas measured. The available task mix affects this estimate; missing areas remain unknown.

Shared-result disagreement. This aggregate order conflicts with matched benchmark results against 1 other configurations. Different test coverage and opponents affect the fit. Inspect published results before treating the rank as conclusive.

Which sources carry the weight?

Artificial Analysis · 28.0% direct comparison weight · independent of the model provider
Epoch AI · 26.6% direct comparison weight · independent of the model provider
ARC Prize · 25.0% direct comparison weight · independent of the model provider
Scale · 20.4% direct comparison weight · independent of the model provider

Which benchmark families carry the weight?

ARC-AGI · 25.0% direct comparison weight
Humanity’s Last Exam · 23.1% direct comparison weight
GPQA · 8.3% direct comparison weight
Aime · 5.6% direct comparison weight
epoch-game-puzzles · 5.3% direct comparison weight
Other families (11) · 32.7% direct comparison weight

These shares describe direct comparison weight, not fractions of the final score or confidence. Source balancing does not give every family equal weight or remove correlations between overlapping sources.

How sensitive is this score?

The audit removes one source at a time and refits against the same overall reference panel. These published results use the matching dataset audited 2026-10-23.

Both ranks use the same surviving configurations for each removal. The cohort starts with 444 configurations supported in the original audit, including separate effort settings. These are diagnostic positions, not the default leaderboard ranks. Eligibility is not rechecked after removal.

What changes when a source is removed?
Removed source	Rank before → after	Rank movement	Cohort remaining / lost support	Score after (change)
Artificial Analysis	78 → 74	Up 4	203 remaining / 241 lost	44.3 (+3.4 points)
ARC Prize	84 → 100	Down 16	444 remaining / 0 lost	38.5 (-2.3 points)
Datacurve	84 → 90	Down 6	444 remaining / 0 lost	36.0 (-4.8 points)
LiveCodeBench	84 → 84	No movement	444 remaining / 0 lost	40.8 (-0.0 points)
Scale	84 → 98	Down 14	444 remaining / 0 lost	37.9 (-2.9 points)
Terminal-Bench	84 → 87	Down 3	444 remaining / 0 lost	40.5 (-0.3 points)
Cognition	84 → 91	Down 7	444 remaining / 0 lost	38.8 (-2.0 points)
LiveBench	84 → 87	Down 3	443 remaining / 1 lost	41.0 (+0.2 points)
SWE-rebench	84 → 85	Down 1	444 remaining / 0 lost	40.6 (-0.2 points)
Epoch AI	83 → 72	Up 11	431 remaining / 13 lost	48.0 (+7.2 points)
Vals AI	84 → 86	Down 2	443 remaining / 1 lost	39.3 (-1.5 points)
Provider reports · Anthropic	84 → 84	No movement	444 remaining / 0 lost	40.8 (-0.0 points)
Provider reports · Moonshot AI	84 → 84	No movement	444 remaining / 0 lost	40.3 (-0.5 points)
Provider reports · Meta	84 → 84	No movement	444 remaining / 0 lost	40.8 (-0.0 points)
Provider reports · Mistral	84 → 84	No movement	444 remaining / 0 lost	40.8 (+0.0 points)
Provider reports · OpenAI	84 → 84	No movement	444 remaining / 0 lost	41.0 (+0.2 points)
All provider reports	84 → 84	No movement	444 remaining / 0 lost	40.5 (-0.4 points)

Lost support means a configuration no longer has a comparison path to the complete reference panel. Its rank is unavailable. A surviving path alone does not establish leaderboard eligibility.

Overlapping sources may remain correlated even with source balancing. Leave-one-source-out movement is a sensitivity diagnostic, not a spread of source-specific ranks or a confidence measure. Scoring methodology · Dated configuration audit

Every benchmark family

Removed evidence	Recomputed score	Change
deepswe	36.0	-4.8
hle	36.8	-4.1
arc-agi	37.7	-3.1
frontiercode	38.8	-2.0
enigma-eval	39.2	-1.6
mmmu-pro	42.4	+1.6
ifbench	42.1	+1.3
omniscience	40.0	-0.8
frontiermath	40.1	-0.7
epoch-game-puzzles	40.2	-0.6
simpleqa	41.4	+0.6
terminal-bench	40.3	-0.5
mmlu-pro	41.3	+0.5
swe-atlas-refactoring	40.4	-0.4
livebench-instructions	41.1	+0.3
gmmlu	40.5	-0.3
aime	41.1	+0.3
livebench-coding	40.5	-0.3
apex-agents	40.6	-0.2
harvey	40.6	-0.2
livebench-language	41.0	+0.2
multi-swe-bench	40.6	-0.2
gdp-pdf	41.0	+0.2
swe-rebench	40.6	-0.2
critpt	40.6	-0.2
swe-atlas-qna	40.6	-0.2
briefcase	40.7	-0.1
automationbench	40.7	-0.1
arc-agi-3	40.7	-0.1
vibe-code	40.7	-0.1
terminal-bench-science	40.9	+0.1
itbench	40.7	-0.1
vals-legal-research	40.7	-0.1
livebench-reasoning	40.7	-0.1
swe-atlas-test-writing	40.7	-0.1
gdpval	40.7	-0.1
vals-finance-agent	40.8	-0.1
livebench-data	40.9	+0.1
gpqa	40.9	+0.1
analyst-agent	40.9	+0.1
tau-bench	40.9	+0.0
vals-tax-agent	40.8	-0.0
vals-code-migration	40.8	-0.0
enterprise-ops	40.8	-0.0
vals-excel-modeling	40.8	-0.0
aa-lcr	40.8	+0.0
livebench-math	40.8	-0.0
math500	40.8	-0.0
livecodebench	40.8	-0.0
swe-bench-pro	40.8	+0.0

What evidence is missing from the fit?

Across all collected settings for this model: 120 observations, 70 contributing, 0 matched without graph weight and 50 excluded. Counts do not establish rank eligibility.

3 · Benchmark family or source protocol has not been reviewed
2 · Effort is not identified
4 · Fewer than two distinct model identities remain in this evaluation
27 · Cited comparator requires original evaluation provenance; not a new matched run.
6 · Supporting evidence outside the reviewed capability core
7 · No reviewed effort for this catalog observation; any separately collected effort measurement is counted separately
1 · Microsoft report Table 11 and section 4.1: MAI Terminal-Bench removes timeouts and uses a minimal ReAct harness; comparator values are cited from other model releases. A common evaluation protocol is not established.

Inspect every observation and exclusion →

Benchmark results

These are published benchmark scores. Supported effort estimates also appear on the Capability leaderboard. Results from different harnesses occupy different rows. Missing cells stay unknown; no scores are copied between effort levels.

Benchmark / source	Low	Medium	High	Max	Thinking	Thinking 32K	Thinking 64K	Non-reasoning	Unspecified
APEX-Agents-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	33.04% Reported settings Claude Opus 4.6 (Max); Model source: https://artificialanalysis.ai/models/claude-opus-4-6-adaptive	Not reported	Not reported	Not reported	Not reported	Not reported
AA-LCR v1.1Artificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	78% Reported settings Claude Opus 4.6 (Max); Model source: https://artificialanalysis.ai/models/claude-opus-4-6-adaptive	Not reported	Not reported	Not reported	67% Reported settings Claude Opus 4.6 (Non-reasoning, High); Model source: https://artificialanalysis.ai/models/claude-opus-4-6	Not reported
CritPtArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark developed by Argonne and UIUC, with contributions from 60+ researchers globally. CritPt is under review by its authors following external feedback · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	12.57% Reported settings Claude Opus 4.6 (Max); Model source: https://artificialanalysis.ai/models/claude-opus-4-6-adaptive	Not reported	Not reported	Not reported	2.84% Reported settings Claude Opus 4.6 (Non-reasoning, High); Model source: https://artificialanalysis.ai/models/claude-opus-4-6	Not reported
Global-MMLU-LiteArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	92.15% Reported settings Claude Opus 4.6 (Max); Model source: https://artificialanalysis.ai/models/claude-opus-4-6-adaptive	Not reported	Not reported	Not reported	90.64% Reported settings Claude Opus 4.6 (Non-reasoning, High); Model source: https://artificialanalysis.ai/models/claude-opus-4-6	Not reported
GPQA DiamondArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	89.6% Reported settings Claude Opus 4.6 (Max); Model source: https://artificialanalysis.ai/models/claude-opus-4-6-adaptive	Not reported	Not reported	Not reported	84.04% Reported settings Claude Opus 4.6 (Non-reasoning, High); Model source: https://artificialanalysis.ai/models/claude-opus-4-6	Not reported
Humanity's Last ExamArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	39.94% Reported settings Claude Opus 4.6 (Max); Model source: https://artificialanalysis.ai/models/claude-opus-4-6-adaptive	Not reported	Not reported	Not reported	19.09% Reported settings Claude Opus 4.6 (Non-reasoning, High); Model source: https://artificialanalysis.ai/models/claude-opus-4-6	Not reported
IFBenchArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	53.13% Reported settings Claude Opus 4.6 (Max); Model source: https://artificialanalysis.ai/models/claude-opus-4-6-adaptive	Not reported	Not reported	Not reported	44.63% Reported settings Claude Opus 4.6 (Non-reasoning, High); Model source: https://artificialanalysis.ai/models/claude-opus-4-6	Not reported
MMMU-ProArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	75.43% Reported settings Claude Opus 4.6 (Max); Model source: https://artificialanalysis.ai/models/claude-opus-4-6-adaptive	Not reported	Not reported	Not reported	72.54% Reported settings Claude Opus 4.6 (Non-reasoning, High); Model source: https://artificialanalysis.ai/models/claude-opus-4-6	Not reported
AA-Omniscience IndexArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details AA-Omniscience Index (higher is better) measures knowledge reliability and hallucination. It rewards correct answers, penalizes hallucinations, and has no penalty for refusing to answer. Scores range from -100 to 100, where 0 means as many correct as incorrect answers, and negative scores mean more incorrect than correct. · Evaluation results measured independently by Artificial Analysis Unit: index. Source collected 2026-10-18.	Not reported	Not reported	Not reported	13.67 Reported settings Claude Opus 4.6 (Max); Model source: https://artificialanalysis.ai/models/claude-opus-4-6-adaptive	Not reported	Not reported	Not reported	2.37 Reported settings Claude Opus 4.6 (Non-reasoning, High); Model source: https://artificialanalysis.ai/models/claude-opus-4-6	Not reported
AA-Omniscience AccuracyArtificial Analysis · 6000 questions; 1 attempt; correct-answer share; GPT-5.6 Luna medium grader Benchmark details AA-Omniscience Accuracy (higher is better) measures the proportion of correctly answered questions out of all questions, regardless of whether the model chooses to answer · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	46.98% Reported settings Claude Opus 4.6 (Max); Model source: https://artificialanalysis.ai/models/claude-opus-4-6-adaptive	Not reported	Not reported	Not reported	45.78% Reported settings Claude Opus 4.6 (Non-reasoning, High); Model source: https://artificialanalysis.ai/models/claude-opus-4-6	Not reported
𝜏²-Bench TelecomArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	92.11% Reported settings Claude Opus 4.6 (Max); Model source: https://artificialanalysis.ai/models/claude-opus-4-6-adaptive	Not reported	Not reported	Not reported	84.8% Reported settings Claude Opus 4.6 (Non-reasoning, High); Model source: https://artificialanalysis.ai/models/claude-opus-4-6	Not reported
Terminal-Bench HardArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	46.21% Reported settings Claude Opus 4.6 (Max); Model source: https://artificialanalysis.ai/models/claude-opus-4-6-adaptive	Not reported	Not reported	Not reported	48.48% Reported settings Claude Opus 4.6 (Non-reasoning, High); Model source: https://artificialanalysis.ai/models/claude-opus-4-6	Not reported
ARC-AGI-2ARC Prize · v2 Semi-Private Benchmark details Original public board. Configuration and task coverage may differ; inspect source. Unit: percent. Source collected 2026-10-18.	64.58% Reported settings Claude Opus 4.6 (120K, Low); Source model ID: claude-opus-4-6-thinking-120K-low; 120 tasks	66.25% Reported settings Claude Opus 4.6 (120K, Medium); Source model ID: claude-opus-4-6-thinking-120K-medium; 120 tasks	69.17% Reported settings Claude Opus 4.6 (120K, High); Source model ID: claude-opus-4-6-thinking-120K-high; 120 tasks	68.75% Reported settings Claude Opus 4.6 (120K, Max); Source model ID: claude-opus-4-6-thinking-120K-max; 120 tasks	Not reported	Not reported	Not reported	Not reported	Not reported
SWE-bench Pro publicScale · mini-swe-agent starred rows Benchmark details Original public board. Configuration and task coverage may differ; inspect source. Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	51.9% Reported settings claude-opus-4-6 (thinking)*; 731 tasks; source date 2026-04-08	Not reported	Not reported	Not reported	Not reported
FrontierCode 1.1 · extended-claude-codeCognition · claude-code Benchmark details FrontierCode 1.1; claude-code. Open each result for its exact effort, task subset, metric and protocol details. Unit: percent. Source collected 2026-10-18.	32.93% Reported settings FrontierCode 1.1; extended; 150 tasks; claude-code; source effort low; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate 0.002	38.65% Reported settings FrontierCode 1.1; extended; 150 tasks; claude-code; source effort medium; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate 0.0007	43.67% Reported settings FrontierCode 1.1; extended; 150 tasks; claude-code; source effort high; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate 0.0007	Not reported	Not reported	Not reported	Not reported	Not reported	Not reported
FrontierCode 1.1 · main-claude-codeCognition · claude-code Benchmark details FrontierCode 1.1; claude-code. Open each result for its exact effort, task subset, metric and protocol details. Unit: percent. Source collected 2026-10-18.	18.31% Reported settings FrontierCode 1.1; main; 100 tasks; claude-code; source effort low; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate 0.003	23.67% Reported settings FrontierCode 1.1; main; 100 tasks; claude-code; source effort medium; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate 0.001	26.64% Reported settings FrontierCode 1.1; main; 100 tasks; claude-code; source effort high; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate 0.001	Not reported	Not reported	Not reported	Not reported	Not reported	Not reported
HLE · text only · Scale official evaluationScale · Scale official evaluation Benchmark details Scale text-only public HLE subset; shares the full HLE family budget; automatic answer judging. Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	36.24% Reported settings Scale text-only public HLE subset; shares the full HLE family budget; automatic answer judging.; published row 2026-02-17T17:06:06.000Z; reported margin ±2.03 percentage points; Potential contamination warning: This model was evaluated after the public release of HLE, allowing model builder access to the prompts and solutions.	Not reported	Not reported	Not reported	19.37% Reported settings Scale text-only public HLE subset; shares the full HLE family budget; automatic answer judging.; published row 2026-02-17T17:06:51.000Z; reported margin ±1.67 percentage points; Potential contamination warning: This model was evaluated after the public release of HLE, allowing model builder access to the prompts and solutions.	Not reported
HLE · full · Scale official evaluationScale · Scale official evaluation Benchmark details Scale full public HLE; 2,500 questions; automatic answer judging; temperature 0 when configurable. Public-set contamination caveats retained per result. Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	Not reported	34.44% Reported settings Scale full public HLE; 2,500 questions; automatic answer judging; temperature 0 when configurable. Public-set contamination caveats retained per result.; published row 2026-02-17T17:04:32.000Z; reported margin ±1.86 percentage points; Potential contamination warning: This model was evaluated after the public release of HLE, allowing model builder access to the prompts and solutions.	Not reported	Not reported	Not reported	19% Reported settings Scale full public HLE; 2,500 questions; automatic answer judging; temperature 0 when configurable. Public-set contamination caveats retained per result.; published row 2026-02-17T17:05:29.000Z; reported margin ±1.54 percentage points; Potential contamination warning: This model was evaluated after the public release of HLE, allowing model builder access to the prompts and solutions.	Not reported
LiveBench · AMPS Hard · 2026-06-25LiveBench · LiveBench official task evaluation Benchmark details LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only. Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	97% Reported settings LiveBench official task evaluation; livebench-2026-06-25-AMPS_Hard; claude-opus-4-6-thinking-auto-high-effort; checkpoint=claude-opus-4-6	Not reported	Not reported	Not reported	Not reported	Not reported	Not reported
LiveBench · code completion · 2026-06-25LiveBench · LiveBench official task evaluation Benchmark details LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only. Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	76.09% Reported settings LiveBench official task evaluation; livebench-2026-06-25-code_completion; claude-opus-4-6-thinking-auto-high-effort; checkpoint=claude-opus-4-6	Not reported	Not reported	Not reported	Not reported	Not reported	Not reported
LiveBench · code generation · 2026-06-25LiveBench · LiveBench official task evaluation Benchmark details LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only. Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	80.28% Reported settings LiveBench official task evaluation; livebench-2026-06-25-code_generation; claude-opus-4-6-thinking-auto-high-effort; checkpoint=claude-opus-4-6	Not reported	Not reported	Not reported	Not reported	Not reported	Not reported
LiveBench · connections · 2026-06-25LiveBench · LiveBench official task evaluation Benchmark details LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only. Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	99.33% Reported settings LiveBench official task evaluation; livebench-2026-06-25-connections; claude-opus-4-6-thinking-auto-high-effort; checkpoint=claude-opus-4-6	Not reported	Not reported	Not reported	Not reported	Not reported	Not reported
LiveBench · consecutive events · 2026-06-25LiveBench · LiveBench official task evaluation Benchmark details LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only. Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	63.47% Reported settings LiveBench official task evaluation; livebench-2026-06-25-consecutive_events; claude-opus-4-6-thinking-auto-high-effort; checkpoint=claude-opus-4-6	Not reported	Not reported	Not reported	Not reported	Not reported	Not reported
LiveBench · integrals with game · 2026-06-25LiveBench · LiveBench official task evaluation Benchmark details LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only. Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	73% Reported settings LiveBench official task evaluation; livebench-2026-06-25-integrals_with_game; claude-opus-4-6-thinking-auto-high-effort; checkpoint=claude-opus-4-6	Not reported	Not reported	Not reported	Not reported	Not reported	Not reported
LiveBench · javascript · 2026-06-25LiveBench · LiveBench Mini-SWE-Agent, 250-step limit Benchmark details LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only. Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	63.64% Reported settings LiveBench Mini-SWE-Agent, 250-step limit; livebench-2026-06-25-javascript; claude-opus-4-6-thinking-auto-high-effort; checkpoint=claude-opus-4-6	Not reported	Not reported	Not reported	Not reported	Not reported	Not reported

What these comparisons mean

Each source reports its own effort labels. Matching labels do not establish equal compute budgets or identical fallback behavior. Open Reported settings to inspect the exact model name, agent and task coverage. Multiple results at one level remain visible rather than selecting the best. Source-reported None or Non-reasoning labels are not treated as evidence that an API supports those settings.

These measurements use their own dated source collection. The leaderboard fits identified configurations jointly; it contains no pooled model ratings. Explicit thinking budgets are separate configurations. Read the methodology.

Download all effort benchmark results · 8,916 results across 259 models · Collected 2026-10-18

Claude Opus 4.6

Compare with

One configuration, seven capabilities

Agentic

Hard reasoning

Coding

Human pref

Knowledge

Multimodal

Long context

Capability score evidence

Benchmark results