Which is better, Claude Opus 4.6 or Qwen3.6 Plus?

Across 61 shared benchmarks tracked on Vector Wire, Claude Opus 4.6 leads on 48 and Qwen3.6 Plus 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 Qwen3.6 Plus

AnthropicvsAlibaba61 shared benchmarks48–12 head-to-head

Benchmark	Claude Opus 4.6	Qwen3.6 Plus
AA Agentic Index	67.6	29
AA Intelligence	44.9	40.5
AA-LCR	74.3	72.3
AA-Omniscience	13.7	2.6
AgentWorldBench - Android	61.7	57.6
AgentWorldBench - MCP	69.9	55.3
AgentWorldBench - OS	70.2	60.3
AgentWorldBench - Overall	57.8	50.8
AgentWorldBench - Search	29.3	21.9
AgentWorldBench - SWE	64.5	59.1
AgentWorldBench - Terminal	57.5	50.6
AgentWorldBench - Web	51.4	50.8
aime_2026	96.7	95.3
arena_elo	1497	1444
arena_text_factuality	1487	1438
Artificial Analysis Coding Index	48.1	54.5
coding_arena_elo	1545	1460
critpt	12.6	2.9
ERQA	40.8	65.7
frontiermath_tier_4	22.9	8.3
gdpval	55.9	32
GPQA Diamond	91.3	90.4
HLE	53.1	28.8
HLE (with tools)	62.7	50.6
hmmt_feb_2026	96.2	87.8
hmmt_nov_2025	96.3	94.6
IFBench	62.5	75.2
imo_answer_bench	75.3	83.8
livebench	76.3	70.8
livebench_agentic_coding	49	41.4
livebench_coding	78.2	78.2
livebench_data_analysis	69.9	69.9
livebench_instruction_following	63.3	58.3
livebench_language	83.3	75
livebench_math	89.3	83.7
livebench_reasoning	88.7	75.8
LiveCodeBench v6	88.8	87.1
mathvision	71.2	88
MCP Atlas	76.8	74.1
mcpmark	56.7	48.2
MMLU-Pro	89.1	88.5
mmmlu	91.1	89.5
MMMU-Pro	77.3	78.8
nl2repo	49.8	37.9
OmniDocBench 1.5	86.6	91.2
OmniScience Accuracy	47	26.4
OmniScience Non-Hallucination	37.2	68
OSWorld-Verified	72.7	62.5
RealWorldQA	73.9	85.4
scicode	52	40.7
screenspot_pro_no_tools	57.7	68.2
simpleqa_verified	46.5	49.1
SWE-bench Multilingual	77.8	73.8
SWE-bench Pro	57.3	56.6
SWE-bench Verified	80.8	78.8
Terminal-Bench 2.0	65.4	61.6
Terminal-Bench Hard	65.4	43.9
tool_decathlon	47.2	39.8
toolathlon	56.8	39.8
τ²-Bench Telecom (AA run)	99.3	97.7
τ³-Bench	72.4	70.7

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.