Which is better, Claude Opus 4.6 or Claude Opus 4.7?

Across 69 shared benchmarks tracked on Vector Wire, Claude Opus 4.6 leads on 20 and Claude Opus 4.7 leads on 48 — Claude Opus 4.7 leads on more. Strengths differ by axis; the table below shows every head-to-head number with provenance.

SEEDRefresh Models —Deals —Regulatory —Sources —

Claude Opus 4.6 vs Claude Opus 4.7

AnthropicvsAnthropic69 shared benchmarks20–48 head-to-head

Benchmark	Claude Opus 4.6	Claude Opus 4.7
AA Agentic Index	67.6	64.6
AA Intelligence	44.9	55
AA-LCR	74.3	75.3
AA-Omniscience	13.7	27.3
aime_2026	96.7	95.8
arc_agi_1	93	92
arc_agi_3	0.5	0.2
ARC-AGI-2	68.8	75.8
arena_elo	1497	1494
arena_text_factuality	1487	1474
arena_vision	1300	1304
Artificial Analysis Coding Index	48.1	73.6
BigLaw Bench	90.2	90.9
browsecomp	86.8	79.8
charxiv_reasoning_with_tools	84.7	91
charxiv_rq	69.1	82.1
coding_arena_elo	1545	1557
critpt	12.6	12
cybergym	73.8	73.8
finance_agent	76.7	71.5
frontiermath_tier_4	22.9	22.9
gdpval	55.9	58.6
GDPval-AA (Elo)	1619	1753
GPQA Diamond	91.3	94.2
HiL-Bench	38.3	41.7
HLE	53.1	54.7
HLE (with tools)	62.7	54.7
hmmt_feb_2026	96.2	93.9
IFBench	62.5	58.6
Legal Agent Benchmark	4.2	7.1
livebench	76.3	76.9
livebench_agentic_coding	49	50.7
livebench_coding	78.2	82.1
livebench_data_analysis	69.9	78.3
livebench_instruction_following	63.3	66.7
livebench_language	83.3	77.9
livebench_math	89.3	92.8
livebench_reasoning	88.7	87.2
MCP Atlas	76.8	79.1
mmmlu	91.1	91.5
MMMU-Pro	77.3	78.8
mrcr_v2_8needle_128k_average	84	59.3
officeqa	73.5	86.3
officeqa_pro	57.1	80.6
OmniScience Accuracy	47	48.9
OmniScience Non-Hallucination	37.2	57.7
OSWorld-Verified	72.7	82.8
scicode	52	54.5
screenspot_pro_no_tools	57.7	79.5
screenspot_pro_with_tools	83.1	87.6
simplebench	67.6	61.7
simpleqa_verified	46.5	50.6
swe_bench_multimodal	27.1	34.5
SWE-bench Multilingual	77.8	80.5
SWE-bench Pro	57.3	64.3
SWE-bench Verified	80.8	87.6
Terminal-Bench 2.0	65.4	69.4
Terminal-Bench Hard	65.4	54.5
toolathlon	56.8	59.3
Toolathlon Avg turns	16.9	25.9
Toolathlon Pass@∞	47.2	52.8
usamo_2026	66.2	69.3
vectara_answer_rate	99.8	98
vectara_avg_summary_length	137.6	149.1
vectara_factual_consistency	87.8	88
vectara_hallucination_rate ↓	12.2	12
XBOW visual-acuity benchmark	54.5	98.5
τ²-Bench Telecom (AA run)	99.3	88.6
τ³-Bench	72.4	28.9

Best tracked score per model per benchmark (default configuration; source-attributed). ↓ marks lower-is-better metrics. Open either model for its full surface, provenance and pricing.