Anthropicproprietary
Released: 2026-09-22Artificial Analysis

Claude Opus 5.5

Anthropic Opus 5.5 high-capacity reasoning model engineered for long-running autonomous knowledge and software refactoring.

Intelligence Index
59.1/ 100
Artificial Analysis Intelligence Index (independent composite)
Throughput Speed
114tok/sec
Output streaming throughput
Time to First Answer
405s
Time to first answer token, including any thinking
Cost to Run the Index
10.04USD
Artificial Analysis cost of the full Index run

Measured results · Artificial Analysis

Independent evaluations of Claude Opus 5.5 (Max, Default Fallback configuration). Values are reported as published; nothing here is estimated or derived. Snapshot 2026-11-15.

Intelligence Index
59.1
Cost to run the Index
$10.04
Time per task
25.4 min
Omniscience index
48.1
Reasoning model
Yes
Cached-input price
$0.2 / 1M

Agentic work

Multimodal & science

Reasoning-effort variants evaluated

VariantIntelligence IndexCost to run IndexOutput speed (tok/s)
Max, Default Fallback59.1$10.04113.5
Xhigh, Default Fallback57.4$6.05109.3
High, Default Fallback51.8$3.46107.0
Medium, Default Fallback48.2$2.45105.4
Low, Default Fallback34.8$0.98103.8
Full analysis on Artificial Analysis

Headline Evaluation Breakdown

Six independently measured evaluations (Artificial Analysis)
Reasoning · CritPt v2 (Hard)68.0%
Coding & agents · Terminal-Bench 4.059.6%
Long context · AA-LCR84.7%
Knowledge · Humanity's Last Exam61.4%

Sourced Benchmark Evaluations (23)

Verified performance across authoritative benchmarks with provenance tracking.
BenchmarkStatusScoreTrustDateSource TypeProvenance
AA-Omniscience (Non-Hallucination)Active49.7%
93
2026-11-15independentSource ↗
VIALS-AAActive61.5%
93
2026-11-15independentSource ↗
ITBench-AAActive38.2%
93
2026-11-15independentSource ↗
Harvey LAB-AAActive4.2%
93
2026-11-15independentSource ↗
AA-AnalystAgentActive56.3%
93
2026-11-15independentSource ↗
AutomationBench-AAActive69.5%
93
2026-11-15independentSource ↗
Multimodal Knowledge Agent (MKA)Active53%
93
2026-11-15independentSource ↗
MLCR-AA (Medical Long Context Reasoning)Active27.8%
93
2026-11-15independentSource ↗
GDP.pdfActive26.2%
93
2026-11-15independentSource ↗
FrontierCode-AA 2.0Active46.4%
93
2026-11-15independentSource ↗
HealthBench ProfessionalActive59.9%
93
2026-11-15independentSource ↗
HealthBenchActive57%
93
2026-11-15independentSource ↗
FrontierCodeActive52.2%
98
2026-11-15independentSource ↗
AA-Omniscience (Accuracy)Active65.5%
93
2026-11-15independentSource ↗
HLE-DiamondActive67.9%
93
2026-11-15independentSource ↗
CritPt v2 (Hard)Active68%
93
2026-11-15independentSource ↗
CritPtActive31.7%
93
2026-11-15independentSource ↗
AA-LCR (Long Context Reasoning)Active84.7%
93
2026-11-15independentSource ↗
Terminal-Bench-Science 0.1Active59%
93
2026-11-15independentSource ↗
Terminal-Bench 4.0Active59.6%
93
2026-11-15independentSource ↗
SciCodeActive66.9%
93
2026-11-15independentSource ↗
MMMU-ProActive87.7%
93
2026-11-15independentSource ↗
Humanity's Last ExamActive61.4%
93
2026-11-15independentSource ↗
Anthropic Python SDKAPI Ready
import anthropic

client = anthropic.Anthropic()
message = client.messages.create(
    model="claude-opus-5-5",
    max_tokens=2048,
    temperature=0,
    messages=[{"role": "user", "content": "Execute SWE-bench verified repository fix"}]
)
print(message.content[0].text)

Evaluation & Sourcing Notes

Scores listed for Claude Opus 5.5 represent verified evaluations extracted from official research papers, independent evaluation suites (HELM, LMSYS, OpenCompass, LiveBench), and verified audit reports.

All benchmarks marked as saturated or deprecated reflect historical performance where the benchmark no longer provides active discriminative power.