Claude Opus 4.8 (Adaptive Reasoning, Max Effort)

May 28, 2026Last synced Aug 1, 2026, 12:19 AM1000K context

Human Frontier

83.9

Human-calibrated frontier signal, backed by Arena-style preference evidence and separate from raw AgMoBench benchmark composite scores.

Blended Price

$10.00/M

Input Price

$5.00/M

Output Price

$25.00/M

Speed

—

TTFT

—

Benchmark Scores

Intelligence Index

Performance Radar

Claude Opus 4.8 (Adaptive Reasoning, Max Effort)

Field Average

External Benchmarks

AA-Omniscience Accuracy

How Claude Opus 4.8 (Adaptive Reasoning, Max Effort) Compares

Axes

X Axis

Y Axis

Bubble Size

Filters

Blended Price (USD)$0.00 – $37.5

AgMoBench Overall≥ 1.8

Providers

ai2ai21-labsalibabaanthropicarceeawsazurebaidu

Show quadrants

anthropic

openai

mistral

azure

cohere

meta

ai21-labs

inclusionai

google

ai2

ibm

alibaba

upstage

nvidia

aws

kimi

nous-research

thinking-machines

minimax

stepfun

tencent

deepseek

zai

multiversecomputing

xiaomi

xai

deepcogito

sarvam

kwaikat

baidu

swiss-ai-initiative

bytedance_seed

inception

reka-ai

arcee

perplexity

Bubble size = Context Window

8191.00

2000000.00

Compare with other models

ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · Claude Opus 4.8 (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:anthropic-opus-4-8-high

View run source

Source metadata

{
  "provider": "Anthropic",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "anthropic-opus-4-8",
  "modelNotes": "Evaluation conducted with 'high' output effort.",
  "modelsEtag": "W/\"2fc19b6caba8b52765ca17eab04fc437\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"9485c7ef206d9b12cdfdec6a631df56b\"",
  "upstreamModelId": "anthropic-opus-4-8-high",
  "environmentCount": 25,
  "modelReleaseDate": "2026-06-01T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 10000,
  "providerModelIdPublished": false
}

Evidence observed 2026-07-31

1.5 / 100

ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · Claude Opus 4.8 (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:anthropic-opus-4-8-high

View run source

Source metadata

{
  "provider": "Anthropic",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "anthropic-opus-4-8",
  "modelNotes": "Evaluation conducted with 'high' output effort.",
  "modelsEtag": "W/\"2fc19b6caba8b52765ca17eab04fc437\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"9485c7ef206d9b12cdfdec6a631df56b\"",
  "upstreamModelId": "anthropic-opus-4-8-high",
  "environmentCount": 25,
  "modelReleaseDate": "2026-06-01T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 10000,
  "providerModelIdPublished": false
}

Evidence observed 2026-07-31

1.5 / 100

ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · Claude Opus 4.8 (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:anthropic-opus-4-8-high

View run source

Source metadata

{
  "provider": "Anthropic",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "anthropic-opus-4-8",
  "modelNotes": "Evaluation conducted with 'high' output effort.",
  "modelsEtag": "W/\"cfc9b87c645d3ef2103e359fd0f9f521\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"0d65f6e0efe0d5cae15e8e2fe5a0994b\"",
  "upstreamModelId": "anthropic-opus-4-8-high",
  "environmentCount": 25,
  "modelReleaseDate": "2026-06-01T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 10000,
  "providerModelIdPublished": false
}

Evidence observed 2026-07-30

1.5 / 100

ARC-AGI-3 Public Demoarc_agi

v3 · public-demo · Claude Opus 4.8 (High) · effort high · 1 attempt · 5x-human-baseline-median-per-level · arc-agi-3:public-demo · record v3_Semi_Private:anthropic-opus-4-8-high

View run source

Source metadata

{
  "provider": "Anthropic",
  "modelType": "CoT",
  "modelsUrl": "https://arcprize.org/media/data/models.json",
  "modelGroup": "anthropic-opus-4-8",
  "modelNotes": "Evaluation conducted with 'high' output effort.",
  "modelsEtag": "W/\"bbdf5725ef7a703a7f0970e913262b6f\"",
  "scoreMetric": "relative-human-action-efficiency",
  "evaluationsUrl": "https://arcprize.org/media/data/evaluations.json",
  "evaluationsEtag": "W/\"6529094085f9ca7ce5f3c2e1957d3b9c\"",
  "upstreamModelId": "anthropic-opus-4-8-high",
  "environmentCount": 25,
  "modelReleaseDate": "2026-06-01T00:00:00.000Z",
  "upstreamDatasetId": "v3_Semi_Private",
  "actionBudgetPolicy": "5x-human-baseline-median-per-level",
  "technicalReportUrl": "https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf",
  "catalogMatchAliases": [],
  "agentHarnessPublished": false,
  "verificationAuthority": "ARC Prize Foundation",
  "upstreamRunIdPublished": false,
  "actualEvaluationCostUsd": 10000,
  "providerModelIdPublished": false
}

Evidence observed 2026-07-24

1.5 / 100

EnigmaEvalseal

vlive-leaderboard · enigma_eval · claude-opus-4-8-xhigh · effort xhigh · 1 attempt · pass@1 · scale-seal · provider claude-opus-4-8-xhigh · record enigma_eval:claude-opus-4-8-xhigh

View run source

Source metadata

{
  "rank": 4,
  "company": "anthropic",
  "benchmarkId": "enigma_eval",
  "identityNote": "Scale SEAL published configuration label preserved",
  "benchmarkName": "EnigmaEval"
}

Evidence observed 2026-08-01

23.5

EnigmaEvalseal

vlive-leaderboard · enigma_eval · claude-opus-4-8-xhigh · effort xhigh · 1 attempt · pass@1 · scale-seal · provider claude-opus-4-8-xhigh · record enigma_eval:claude-opus-4-8-xhigh

View run source

Source metadata

{
  "rank": 4,
  "company": "anthropic",
  "benchmarkId": "enigma_eval",
  "identityNote": "Scale SEAL published configuration label preserved",
  "benchmarkName": "EnigmaEval"
}

Evidence observed 2026-07-31

23.5

EnigmaEvalseal

vlive-leaderboard · enigma_eval · claude-opus-4-8-xhigh · effort xhigh · 1 attempt · pass@1 · scale-seal · provider claude-opus-4-8-xhigh · record enigma_eval:claude-opus-4-8-xhigh

View run source

Source metadata

{
  "rank": 4,
  "company": "anthropic",
  "benchmarkId": "enigma_eval",
  "identityNote": "Scale SEAL published configuration label preserved",
  "benchmarkName": "EnigmaEval"
}

Evidence observed 2026-07-30

23.5

EnigmaEvalseal

vlive-leaderboard · enigma_eval · claude-opus-4-8-xhigh · effort xhigh · 1 attempt · pass@1 · scale-seal · provider claude-opus-4-8-xhigh · record enigma_eval:claude-opus-4-8-xhigh

View run source

Source metadata

{
  "rank": 4,
  "company": "anthropic",
  "benchmarkId": "enigma_eval",
  "identityNote": "Scale SEAL published configuration label preserved",
  "benchmarkName": "EnigmaEval"
}

Evidence observed 2026-07-29

23.5

Senior SWE-Bench v2026.06senior_swe_bench

v2026.06 · official-100-task · Claude Opus 4.8 (max) · harness mini-swe-agent · effort max · 1 attempt · pass@1 · harbor:snorkel-ai/senior-swe-bench-v2026.06 · provider anthropic/claude-opus-4-8 · record opus-48-mini-swe · run 4ee23dca8f25cd97660a6a5e6781ba4c8a820780bc8018dd43fcf22a96c06b11

View run source

Source metadata

{
  "taskCount": 100,
  "agentGroup": "mini-swe-agent:claude-opus-4-8",
  "basicPassAt1": 39.77272727272727,
  "datasetPackage": "snorkel-ai/senior-swe-bench-v2026.06",
  "publicTaskCount": 50,
  "tastefulPassAt1": 25,
  "agentDescription": "Claude Opus 4.8 via mini-swe-agent harness (Portkey→Bedrock)",
  "privateTaskCount": 50,
  "upstreamProvider": "anthropic",
  "upstreamRunCount": 80,
  "averageAgentSteps": 132.36363636363637,
  "averageOutputTokens": 115169.26136363637,
  "averageTotalCostUsd": 8.03432186590909,
  "averageOutputCostUsd": 2.879231534090909
}

Evidence observed 2026-06-30

25.0 / 100

SWE-rebenchswe_rebench

vcontinuous-leaderboard · tasks-created:2026-03-01:2026-05-15 · Claude Opus 4.8-xhigh · harness swe-rebench-react@tools · effort xhigh · 5 attempts · pass@1 · swe-rebench:standardized:128k-or-model-limit · provider Claude Opus 4.8-xhigh__tools · record Claude Opus 4.8-xhigh__tools:1772323200000:1778803200000

View run source

Source metadata

{
  "passAt5": 67.27272727272727,
  "developer": "anthropic",
  "instanceType": "model",
  "contextPolicy": "128k unless the model supports less",
  "attemptsPerTask": 5,
  "totalTokenUsage": 2479387.470909091,
  "modelReleaseDate": "2026-05-28",
  "scoreStandardError": 1.199173268933903,
  "cachedTokenPercentage": 95.27838279131717,
  "modelReleaseTimestamp": 1779926400000,
  "standardContextTokens": 128000,
  "averageInstanceCostUsd": 2.0202793086363635,
  "potentialContamination": true,
  "aggregateTaskRangeTimestamp": {
    "to": 1778803200000,
    "from": 1772323200000
  },
  "publishedTaskRangeTimestamp": {
    "to": 1778803200000,
    "from": 1772323200000
  }
}

Evidence observed 2026-05-15

56.5 / 100

Terminal-Bench 2.1terminal_bench

v2.1 · official-leaderboard · Opus 4.8 (high) · harness claude-code@2.1.205 · effort high · 5 attempts · harbor:terminal-bench/terminal-bench-2-1 · provider anthropic/claude-opus-4-8 · record leaderboard/submissions/2026-07-09-anthropic-claude-opus-4-8-high-claude-code.json · run a3019ec2-bc78-5ff6-9cae-d22d62470515

View run source

Source metadata

{
  "passAt2": 0.8652,
  "passAt3": 0.9011,
  "passAt4": 0.9258,
  "passAt5": 0.9438,
  "taskCount": 89,
  "actualTokens": {
    "output": 8089069,
    "cachedInput": 161931526,
    "uncachedInput": 12873472
  },
  "datasetPackage": "terminal-bench/terminal-bench-2-1",
  "leaderboardUrl": "https://www.tbench.ai/leaderboard/terminal-bench/2.1",
  "pullRequestUrl": "https://github.com/harbor-framework/terminal-bench-2-1/pull/92",
  "agentDisplayUrl": "https://claude.com/claude-code",
  "modelDisplayUrl": "https://www.anthropic.com/news/claude-opus-4-8",
  "actualTrialCount": 445,
  "agentOrganization": "Anthropic",
  "modelOrganization": "Anthropic",
  "actualTotalCostUsd": 286.94,
  "rewardHacksPercent": 0,
  "accuracyStandardError": 1.31,
  "submissionDownloadUrl": "https://raw.githubusercontent.com/harbor-framework/terminal-bench-2-1/main/leaderboard/submissions/2026-07-09-anthropic-claude-opus-4-8-high-claude-code.json",
  "disqualifiedTrialCount": 0,
  "expectedAttemptsPerTask": 5,
  "sourceFilterReasoningEffort": "high",
  "actualAverageTrialDurationSeconds": 551.4
}

Evidence observed 2026-07-09

78.9 / 100