Which is better, Gemma 4 26B A4B or Qwen3.6 Plus?

Across 52 shared benchmarks tracked on Vector Wire, Gemma 4 26B A4B leads on 4 and Qwen3.6 Plus leads on 48 — Qwen3.6 Plus leads on more. Strengths differ by axis; the table below shows every head-to-head number with provenance.

SEEDRefresh Models —Deals —Regulatory —Sources —

Gemma 4 26B A4B vs Qwen3.6 Plus

GooglevsAlibaba52 shared benchmarks4–48 head-to-head

Benchmark	Gemma 4 26B A4B	Qwen3.6 Plus
AA Agentic Index	32.1	29
AA Intelligence	26.1	40.5
AA-LCR	61.7	72.3
AA-Omniscience	-50.8	2.6
AI2D	88.3	94.4
aime_2026	88.3	95.3
arena_elo	1438	1444
arena_text_factuality	1445	1438
Artificial Analysis Coding Index	39.3	54.5
C-Eval	82.5	93.3
Claw Eval (pass@3)	28	58.7
coding_arena_elo	1361	1460
critpt	0	2.9
DeepPlanning	16.2	41.5
gdpval	25.7	32
GPQA Diamond	82.3	90.4
HLE	19.3	28.8
HLE (with tools)	17.2	50.6
hmmt_feb_2026	79	87.8
hmmt_nov_2025	87.5	94.6
IFBench	72.4	75.2
ifeval	96.4	94.3
imo_answer_bench	74.3	83.8
LiveCodeBench v6	77.1	87.1
mathvision	82.4	88
MCP Atlas	50	74.1
mcpmark	14.2	48.2
mmlu_redux	92.7	94.5
MMLU-Pro	85.2	88.5
mmmlu	86.3	89.5
MMMU	78.4	86
MMMU-Pro	73.8	78.8
nl2repo	11.6	37.9
OmniDocBench 1.5	74.4	91.2
OmniScience Accuracy	19.1	26.4
OmniScience Non-Hallucination	13.6	68
RealWorldQA	72.2	85.4
scicode	40.3	40.7
SimpleVQA	52.2	0.7
supergpqa	61.4	71.6
SWE-bench Multilingual	43.4	73.8
SWE-bench Pro	13.8	56.6
SWE-bench Verified	57.4	78.8
TauBench V3 - Banking	12	20.8
Terminal-Bench 2.0	34.2	61.6
Terminal-Bench 2.1	39	61.4
Terminal-Bench Hard	25	43.9
tool_decathlon	12	39.8
VITA-Bench	36.9	44.3
WideSearch	38.3	74.3
τ²-Bench Telecom (AA run)	43.6	97.7
τ³-Bench	59	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.