Which is better, Qwen3.5 397B A17B or Qwen3.7 Plus Preview?

Across 50 shared benchmarks tracked on Vector Wire, Qwen3.5 397B A17B leads on 9 and Qwen3.7 Plus Preview leads on 40 — Qwen3.7 Plus Preview leads on more. Strengths differ by axis; the table below shows every head-to-head number with provenance.

SEEDRefresh Models —Deals —Regulatory —Sources —

Qwen3.5 397B A17B vs Qwen3.7 Plus Preview

AlibabavsAlibaba50 shared benchmarks9–40 head-to-head

Benchmark	Qwen3.5 397B A17B	Qwen3.7 Plus Preview
AA Agentic Index	53.3	20.8
AA Intelligence	34.3	39.4
AA-LCR	72.7	69
AA-Omniscience	-30.8	2.4
apexAgents	15.3	22.4
arena_elo	1442	1458
arena_text_factuality	1445	1454
arena_vision	1249	1263
Artificial Analysis Coding Index	48.2	55.9
bfcl_v4	72.9	72.9
charxiv_rq	80.8	85.9
Claw Eval (pass@3)	48.1	62.7
critpt	1.7	9.1
DeepPlanning	34.3	62.3
ERQA	67.5	69.8
gdpval	35.8	22.2
GPQA Diamond	89.3	90.3
HLE	29	35.6
hmmt_feb_2026	87.9	92.9
IFBench	78.8	79.1
ifeval	92.6	94.6
imo_answer_bench	80.9	86
include	85.6	83
LiveCodeBench v6	83.6	89.6
mcpmark	46.1	58.7
MLVU	86.7	87.4
mmlu_prox	84.7	85.4
mmlu_redux	94.9	94.5
MMLU-Pro	87.8	88.5
mmmlu	88.5	89
MMMU-Pro	79	80.5
nl2repo	32.2	41.1
OmniScience Accuracy	30.8	22.5
OmniScience Non-Hallucination	17.3	74.5
QwenClawBench	51.8	61.8
RealWorldQA	83.9	86.9
scicode	42	51.3
SimpleVQA	67.1	0.8
supergpqa	70.4	71.4
SWE-bench Multilingual	69.3	75.8
SWE-bench Pro	50.9	57.6
SWE-bench Verified	76.4	77.7
TauBench V3 - Banking	13.4	17.9
Terminal-Bench 2.0	52.5	70.3
Terminal-Bench 2.1	51.3	61
Terminal-Bench Hard	40.9	47
VideoMMMU	84.7	85.4
VITA-Bench	49.7	45.6
τ²-Bench Telecom (AA run)	95.6	93
τ³-Bench	13.4	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.