Which is better, Claude Opus 4.6 or Qwen3.7 Plus Preview?

Across 47 shared benchmarks tracked on Vector Wire, Claude Opus 4.6 leads on 26 and Qwen3.7 Plus Preview leads on 21 — 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 Qwen3.7 Plus Preview

AnthropicvsAlibaba47 shared benchmarks26–21 head-to-head

Benchmark	Claude Opus 4.6	Qwen3.7 Plus Preview
AA Agentic Index	67.6	20.8
AA Intelligence	44.9	39.4
AA-LCR	74.3	69
AA-Omniscience	13.7	2.4
AndroidWorld	62	81
Apex (Pass@1)	34.5	22.7
apexAgents	33	22.4
arena_elo	1497	1458
arena_text_factuality	1487	1454
arena_vision	1300	1263
Artificial Analysis Coding Index	48.1	55.9
baby_vision	14.8	70.4
charxiv_rq	69.1	85.9
ClawEval-MM	54.7	60.1
critpt	12.6	9.1
ERQA	40.8	69.8
gdpval	55.9	22.2
GPQA Diamond	91.3	90.3
HLE	53.1	35.6
hmmt_feb_2026	96.2	92.9
IFBench	62.5	79.1
imo_answer_bench	75.3	86
LiveCodeBench v6	88.8	89.6
mathvision	71.2	90.3
MCP Atlas	76.8	73.2
mcpmark	56.7	58.7
MMLU-Pro	89.1	88.5
mmmlu	91.1	89
MMMU-Pro	77.3	80.5
nl2repo	49.8	41.1
ocrbench_v2	48.4	67.1
OmniDocBench 1.5	86.6	91.4
OmniScience Accuracy	47	22.5
OmniScience Non-Hallucination	37.2	74.5
OSWorld-Verified	72.7	73.3
RealWorldQA	73.9	86.9
scicode	52	51.3
screenspot_pro_no_tools	57.7	79
SWE-bench Multilingual	77.8	75.8
SWE-bench Pro	57.3	57.6
SWE-bench Verified	80.8	77.7
SWE-MM	27.1	30
Terminal Bench 2.1 (Terminus)	78.2	64
Terminal-Bench 2.0	65.4	70.3
Terminal-Bench Hard	65.4	47
τ²-Bench Telecom (AA run)	99.3	93
τ³-Bench	72.4	17.9

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.