Anthropicproprietary
Released: 2026-08-01

Claude Opus 5.5

OpenRouter ranked #11 foundation model processing 2.95T tokens weekly. Deployed by anthropic.

Intelligence Index
82.8/ 100
Calibrated multi-domain composite
Throughput Speed
100tok/sec
Output streaming throughput
Time To First Token
280ms
Initial chunk server latency
LiveBench Contam-Free
—%
September 2026 suite score

Sourced Benchmark Evaluations (60)

Verified performance across authoritative benchmarks with provenance tracking.
BenchmarkStatusScoreTrustDateSource TypeProvenance
ARC-AGI-3Active53.2%
93
2026-09-05independentSource ↗
ARC-AGI-2Nearing Saturation63.8%
75
2026-08-29independentSource ↗
FrontierMathActive22%
98
2026-08-26independentSource ↗
PutnamBenchActive40.9%
90
2026-08-25independentSource ↗
Omni-MATHDeprecated58.2%
10
2026-08-23independentSource ↗
SWE-LancerDeprecated55.7%
20
2026-08-23independentSource ↗
FrontierCodeActive47.5%
98
2026-08-21independentSource ↗
SoohakActive59.8%
98
2026-08-21independentSource ↗
StructFlowBenchNearing Saturation64.7%
75
2026-08-19independentSource ↗
Humanity's Last ExamActive36%
93
2026-08-19independentSource ↗
AIMENearing Saturation74.8%
75
2026-08-19independentSource ↗
OlympiadBenchActive64.5%
80
2026-08-18independentSource ↗
Terminal-BenchActive57.3%
80
2026-08-17independentSource ↗
FireBenchActive64%
95
2026-08-17independentSource ↗
MathArenaActive63%
99
2026-08-17independentSource ↗
SimpleQAActive43.1%
90
2026-08-16independentSource ↗
OpenAI-MRCRActive69.6%
80
2026-08-16independentSource ↗
CodeContestsDeprecated53.2%
10
2026-08-16independentSource ↗
SWE-bench ProActive50.8%
90
2026-08-15independentSource ↗
ComplexBenchNearing Saturation62.5%
80
2026-08-15independentSource ↗
LongBench-ProActive67.9%
80
2026-08-15independentSource ↗
MultiPL-EDeprecated75.3%
10
2026-08-14independentSource ↗
ZeroSCROLLSNearing Saturation66.2%
78
2026-08-14independentSource ↗
InfiniteBenchNearing Saturation70.4%
65
2026-08-13independentSource ↗
LiveCodeBenchSaturated67.2%
39
2026-08-13independentSource ↗
AgentIFNearing Saturation51.8%
80
2026-08-13independentSource ↗
MATHSaturated77.1%
25
2026-08-13independentSource ↗
SCROLLSNearing Saturation66.2%
65
2026-08-12independentSource ↗
MercuryDeprecated72.1%
10
2026-08-12independentSource ↗
MMLU-ProNearing Saturation76.2%
75
2026-08-12independentSource ↗
TabMWPSaturated72.4%
25
2026-08-11independentSource ↗
CyberSecEvalDeprecated73.7%
10
2026-08-11independentSource ↗
RULERNearing Saturation76.2%
84
2026-08-11independentSource ↗
SWE-bench VerifiedSaturated63.9%
25
2026-08-10independentSource ↗
QuALITYNearing Saturation72.9%
78
2026-08-10independentSource ↗
MMLU-ReduxSaturated81.1%
25
2026-08-10independentSource ↗
SWE-benchDeprecated55.7%
10
2026-08-09independentSource ↗
MGSMNearing Saturation74.8%
65
2026-08-09independentSource ↗
LongBenchNearing Saturation71.2%
65
2026-08-09independentSource ↗
QASPERNearing Saturation68.7%
65
2026-08-08independentSource ↗
TruthfulQASaturated72.9%
25
2026-08-08independentSource ↗
AGIEvalSaturated70.3%
25
2026-08-08independentSource ↗
SuperGLUESaturated83.4%
25
2026-08-07independentSource ↗
GSM-HardDeprecated75.6%
10
2026-08-07independentSource ↗
MT-BenchNearing Saturation7.8%
75
2026-08-06independentSource ↗
BIG-Bench HardSaturated80.2%
25
2026-08-06independentSource ↗
CodeXGLUEDeprecated77%
23
2026-08-06independentSource ↗
WinoGrandeSaturated85.1%
28
2026-08-05independentSource ↗
MBPP+Nearing Saturation83.5%
65
2026-08-05independentSource ↗
IFEvalDeprecated74.7%
25
2026-08-05independentSource ↗
GPQANearing Saturation72.9%
78
2026-08-05independentSource ↗
HumanEval+Saturated88.5%
25
2026-08-04independentSource ↗
Needle-in-a-HaystackSaturated96.6%
44
2026-08-04independentSource ↗
CommonsenseQASaturated85.9%
28
2026-08-04independentSource ↗
GSM8KSaturated85%
25
2026-08-03independentSource ↗
GLUESaturated88.3%
28
2026-08-03independentSource ↗
MBPPSaturated86.8%
25
2026-08-03independentSource ↗
HumanEvalSaturated93.4%
25
2026-08-02independentSource ↗
MMLUSaturated84.5%
25
2026-08-02independentSource ↗
ARC-AGISaturated76.9%
25
2026-08-02independentSource ↗
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.