RankingEffort benchmarks

Claude Sonnet 4.5

Compare models

54 measured results · 10 effort settings · 33 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 Sonnet 4.5 · Thinking 32K1 of 7 capabilities supported

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

Agentic
Unknown
Hard reasoning
A18.9Supported
Coding
Unknown
Human pref
Unknown
Knowledge
Unknown
Multimodal
Unknown
Long context
Unknown

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 Sonnet 4.5 · Thinking 32K: 11.1 overall comparison score. Meets the overall evidence requirements.

Hard reasoning only. The available task mix affects this estimate; missing areas remain unknown.

Which sources carry the weight?

ARC Prize · 67.9% direct comparison weight · independent of the model provider
Epoch AI · 32.1% direct comparison weight · independent of the model provider

Which benchmark families carry the weight?

ARC-AGI · 67.9% direct comparison weight
FrontierMath · 19.9% direct comparison weight
GPQA · 12.2% 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	147 → 144	Up 3	203 remaining / 241 lost	10.1 (-1.0 points)
ARC Prize	188 → 258	Down 70	444 remaining / 0 lost	5.5 (-5.6 points)
Datacurve	188 → 191	Down 3	444 remaining / 0 lost	8.6 (-2.5 points)
LiveCodeBench	188 → 183	Up 5	444 remaining / 0 lost	11.3 (+0.2 points)
Scale	188 → 188	No movement	444 remaining / 0 lost	12.6 (+1.5 points)
Terminal-Bench	188 → 188	No movement	444 remaining / 0 lost	10.9 (-0.2 points)
Cognition	188 → 188	No movement	444 remaining / 0 lost	10.7 (-0.4 points)
LiveBench	187 → 185	Up 2	443 remaining / 1 lost	11.4 (+0.3 points)
SWE-rebench	188 → 189	Down 1	444 remaining / 0 lost	11.1 (+0.0 points)
Epoch AI	184 → 170	Up 14	431 remaining / 13 lost	13.4 (+2.3 points)
Vals AI	188 → 191	Down 3	443 remaining / 1 lost	10.4 (-0.7 points)
Provider reports · Anthropic	188 → 188	No movement	444 remaining / 0 lost	11.1 (-0.0 points)
Provider reports · Moonshot AI	188 → 187	Up 1	444 remaining / 0 lost	10.8 (-0.3 points)
Provider reports · Meta	188 → 188	No movement	444 remaining / 0 lost	11.1 (-0.0 points)
Provider reports · Mistral	188 → 188	No movement	444 remaining / 0 lost	11.1 (+0.0 points)
Provider reports · OpenAI	188 → 188	No movement	444 remaining / 0 lost	11.2 (+0.1 points)
All provider reports	188 → 187	Up 1	444 remaining / 0 lost	10.9 (-0.2 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
arc-agi	5.4	-5.7
frontiermath	14.0	+2.9
deepswe	8.6	-2.5
gpqa	8.9	-2.2
epoch-game-puzzles	10.1	-1.0
hle	12.0	+0.9
mmmu-pro	11.9	+0.8
simpleqa	11.9	+0.8
arc-agi-3	10.4	-0.7
frontiercode	10.7	-0.4
mmlu-pro	11.4	+0.3
aime	11.4	+0.3
livebench-instructions	11.3	+0.2
omniscience	10.9	-0.2
livecodebench	11.3	+0.2
multi-swe-bench	10.9	-0.2
harvey	11.0	-0.1
livebench-language	11.2	+0.1
enigma-eval	11.0	-0.1
critpt	11.2	+0.1
swe-atlas-qna	11.0	-0.1
livebench-coding	11.0	-0.1
ifbench	11.0	-0.1
apex-agents	11.2	+0.1
briefcase	11.0	-0.1
vibe-code	11.0	-0.1
automationbench	11.0	-0.1
terminal-bench	11.0	-0.1
tau-bench	11.0	-0.1
livebench-reasoning	11.0	-0.1
terminal-bench-science	11.2	+0.1
aa-lcr	11.0	-0.1
gdp-pdf	11.2	+0.1
analyst-agent	11.1	+0.0
gdpval	11.1	-0.0
swe-atlas-test-writing	11.1	-0.0
swe-rebench	11.1	+0.0
itbench	11.1	-0.0
vals-finance-agent	11.1	-0.0
vals-code-migration	11.1	-0.0
vals-legal-research	11.1	-0.0
vals-tax-agent	11.1	-0.0
vals-excel-modeling	11.1	-0.0
livebench-math	11.1	+0.0
swe-atlas-refactoring	11.1	-0.0
math500	11.1	+0.0
livebench-data	11.1	-0.0
enterprise-ops	11.1	-0.0
swe-bench-pro	11.1	+0.0
gmmlu	11.1	+0.0

What evidence is missing from the fit?

Across all collected settings for this model: 87 observations, 46 contributing, 0 matched without graph weight and 41 excluded. Counts do not establish rank eligibility.

4 · Benchmark family or source protocol has not been reviewed
1 · Original SciCode grading defects are under review; retained as supporting evidence, excluded from aggregate and capability scores. Corrected SciCode-Verified requires a separate protocol review. Sources: https://arxiv.org/abs/2608.04975 and https://artificialanalysis.ai/evaluations/scicode
3 · Effort is not identified
19 · No reviewed effort for this catalog observation; any separately collected effort measurement is counted separately
14 · Supporting evidence outside the reviewed capability core

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	Thinking	Thinking 1K	Thinking 8K	Thinking 16K	Thinking 32K	Thinking 48K	Thinking 59K	Reasoning	Non-reasoning	Unspecified
AA-Briefcase EloArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details AA-Briefcase v1.1 is an agentic knowledge work benchmark developed by Artificial Analysis. AA-Briefcase Elo is a combined metric that aggregates rubric pass rate, analytical quality Elo and presentation Elo · Higher is better · Evaluation results measured independently by Artificial Analysis Unit: elo. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	Not reported	Not reported	Not reported	713.27 Elo Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	Not reported	Not reported
AIME 2025Artificial 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	Not reported	Not reported	Not reported	Not reported	88% Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	37% Reported settings Claude 4.5 Sonnet (Non-reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet	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	Not reported	Not reported	Not reported	Not reported	72.33% Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	54% Reported settings Claude 4.5 Sonnet (Non-reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet	Not reported
AutomationBench-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of task objectives completed with no guardrail violations · Higher is better · Benchmark developed by Zapier · 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	Not reported	Not reported	Not reported	Not reported	14.08% Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	Not reported	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	Not reported	Not reported	Not reported	Not reported	1.14% Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	0% Reported settings Claude 4.5 Sonnet (Non-reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet	Not reported
GDP.pdf: All-pass RateArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of attempts where every atomic criterion passed · 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	Not reported	Not reported	Not reported	Not reported	5.2% Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	Not reported	Not reported
GDPval-AA v2.1 LeaderboardArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Elo rating for performance on real-world work tasks · Anchored to DeepSeek V4.1 Flash (max) at 1600 · Higher is better · Evaluation results measured independently by Artificial Analysis Unit: elo. Source collected 2026-10-18.	Not reported	Not reported	Not reported	Not reported	Not reported	Not reported	Not reported	910.05 Elo Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	Not reported	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	Not reported	Not reported	Not reported	Not reported	89.26% Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	Not reported	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	Not reported	Not reported	Not reported	Not reported	83.43% Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	72.73% Reported settings Claude 4.5 Sonnet (Non-reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet	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	Not reported	Not reported	Not reported	Not reported	17.79% Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	7.23% Reported settings Claude 4.5 Sonnet (Non-reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet	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	Not reported	Not reported	Not reported	Not reported	57.28% Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	42.65% Reported settings Claude 4.5 Sonnet (Non-reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet	Not reported
LiveCodeBenchArtificial 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	Not reported	Not reported	Not reported	Not reported	71.43% Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	59.05% Reported settings Claude 4.5 Sonnet (Non-reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet	Not reported
MMLU-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	Not reported	Not reported	Not reported	Not reported	87.45% Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	85.99% Reported settings Claude 4.5 Sonnet (Non-reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet	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	Not reported	Not reported	Not reported	Not reported	68.73% Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	65.2% Reported settings Claude 4.5 Sonnet (Non-reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet	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	Not reported	Not reported	Not reported	Not reported	-0.08 Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	-9.37 Reported settings Claude 4.5 Sonnet (Non-reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet	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	Not reported	Not reported	Not reported	Not reported	32.87% Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	28.38% Reported settings Claude 4.5 Sonnet (Non-reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet	Not reported
SciCodeArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark curated by scientists across 16 disciplines. SciCode is under review following an independent audit of the dataset · 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	Not reported	Not reported	Not reported	Not reported	45.72% Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	Not reported	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	Not reported	Not reported	Not reported	Not reported	78.07% Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	70.47% Reported settings Claude 4.5 Sonnet (Non-reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet	Not reported
𝜏³-BankingArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark developed by Sierra Research · 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	Not reported	Not reported	Not reported	Not reported	24.54% Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	Not reported	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	Not reported	Not reported	Not reported	Not reported	35.61% Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	28.79% Reported settings Claude 4.5 Sonnet (Non-reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet	Not reported
Terminal-Bench v2.1Artificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark developed by the Laude Institute and open-source contributors · 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	Not reported	Not reported	Not reported	Not reported	55.81% Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	Not reported	Not reported
Terminal-Bench v4.0Artificial Analysis · mini-SWE-agent v2.4.6; 66 tasks; pass@1 averaged over 3 repeats; 500 steps; task timeout up to 8 hours Benchmark details Benchmark developed by the Laude Institute, Stanford, and open-source contributors · 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	Not reported	Not reported	Not reported	Not reported	0% Reported settings Claude 4.5 Sonnet (Reasoning); Model source: https://artificialanalysis.ai/models/claude-4-5-sonnet-thinking	Not reported	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.	Not reported	5.83% Reported settings Claude Sonnet 4.5 (Thinking 1K); Source model ID: claude-sonnet-4-5-20250929-thinking-1k; 120 tasks	6.94% Reported settings Claude Sonnet 4.5 (Thinking 8K); Source model ID: claude-sonnet-4-5-20250929-thinking-8k; 120 tasks	6.94% Reported settings Claude Sonnet 4.5 (Thinking 16K); Source model ID: claude-sonnet-4-5-20250929-thinking-16k; 120 tasks	13.61% Reported settings Claude Sonnet 4.5 (Thinking 32K); Source model ID: claude-sonnet-4-5-20250929-thinking-32k; 120 tasks	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.	14.09% Reported settings Scale text-only public HLE subset; shares the full HLE family budget; automatic answer judging.; published row 2025-10-02T17:29:42.000Z; reported margin ±1.47 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	Not reported	Not reported	Not reported	Not reported	Not reported	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.	13.72% 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 2025-10-02T17:27:45.000Z; reported margin ±1.35 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	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 Sonnet 4.5

Compare with

One configuration, seven capabilities

Agentic

Hard reasoning

Coding

Human pref

Knowledge

Multimodal

Long context

Capability score evidence

Benchmark results