Which is better, Claude Opus 4.7 or Qwen3.7 Max?

Across 42 shared benchmarks tracked on Vector Wire, Claude Opus 4.7 leads on 30 and Qwen3.7 Max leads on 12 — 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 4.7 vs Qwen3.7 Max

AnthropicvsAlibaba42 shared benchmarks30–12 head-to-head

Benchmark	Claude Opus 4.7	Qwen3.7 Max
AA Agentic Index	64.6	30.9
AA Intelligence	55	46.7
AA-LCR	75.3	74.7
AA-Omniscience	27.3	13.5
aime_2026	95.8	97
arena_elo	1494	1474
arena_text_factuality	1474	1479
Artificial Analysis Coding Index	73.6	66
coding_arena_elo	1557	1517
critpt	12	13.4
Finance Agent v2	51.5	48.4
gdpval	58.6	38.5
GPQA Diamond	94.2	92.4
HLE	54.7	41.4
HLE (with tools)	54.7	53.5
hmmt_feb_2026	93.9	97.1
IFBench	58.6	80.5
itbenchSre	46.7	42.5
livebench	76.9	74.3
livebench_agentic_coding	50.7	43.6
livebench_coding	82.1	74.2
livebench_data_analysis	78.3	71.8
livebench_instruction_following	66.7	74
livebench_language	77.9	79.7
livebench_math	92.8	85.3
livebench_reasoning	87.2	83.3
MCP Atlas	79.1	76.4
mmmlu	91.5	90.3
OmniScience Accuracy	48.9	31.1
OmniScience Non-Hallucination	57.7	74.4
scicode	54.5	53.5
simplebench	61.7	70.4
simpleqa_verified	50.6	58.5
SWE-bench Multilingual	80.5	78.3
SWE-bench Pro	64.3	60.6
SWE-bench Verified	87.6	80.4
TauBench V3 - Banking	34.6	11.8
Terminal-Bench 2.0	69.4	69.7
Terminal-Bench 2.1	83.1	75
Terminal-Bench Hard	54.5	50.8
τ²-Bench Telecom (AA run)	88.6	94.7
τ³-Bench	28.9	10.9

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.