Which is better, Claude Opus 4.6 or GPT-5.1?

Across 47 shared benchmarks tracked on Vector Wire, Claude Opus 4.6 leads on 35 and GPT-5.1 leads on 12 — Claude Opus 4.6 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 GPT-5.1

AnthropicvsOpenAI47 shared benchmarks35–12 head-to-head

Benchmark	Claude Opus 4.6	GPT-5.1
AA Agentic Index	67.6	32.2
AA Intelligence	44.9	37.5
AA-LCR	74.3	76.7
AA-Omniscience	13.7	5.4
AIME 2025	99.8	94.2
arc_agi_1	93	72.8
ARC-AGI-2	68.8	17.6
arena_elo	1497	1439
arena_text_factuality	1487	1445
arena_vision	1300	1250
Artificial Analysis Coding Index	48.1	49.4
browsecomp	86.8	50.8
coding_arena_elo	1545	1421
critpt	12.6	4.9
Fortress	20.5	25.7
frontiermath_tier_4	22.9	12.5
gdpval	55.9	25
GPQA Diamond	91.3	88.1
HLE	53.1	28.5
HLE (with tools)	62.7	42.7
hmmt_nov_2025	96.3	91.7
IFBench	62.5	72.9
LiveCodeBench v6	88.8	87
matharena_visual_math_overall	72.3	76.9
MCP Atlas	76.8	50.1
MMLU-Pro	89.1	87
mmmlu	91.1	91
MMMU-Pro	77.3	75.5
multichallenge	56	63.4
MultiNRC	57.1	49
OmniScience Accuracy	47	37.7
OmniScience Non-Hallucination	37.2	48.1
scicode	52	43.3
SEAL VISTA	46.1	43.8
simplebench	67.6	53.2
simpleqa_verified	46.5	48.9
swe_bench_bash	75.6	66
SWE-bench Verified	80.8	76.3
Tau2 retail	91.9	77.9
Terminal-Bench 2.0	65.4	47.6
Terminal-Bench Hard	65.4	45.5
vectara_answer_rate	99.8	100
vectara_avg_summary_length	137.6	254.4
vectara_factual_consistency	87.8	89.1
vectara_hallucination_rate ↓	12.2	12.1
τ²-Bench Telecom (AA run)	99.3	81.9
τ³-Bench	72.4	14

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.