Which is better, Claude Opus 4.6 or Gemini 3 Flash Preview?

Across 53 shared benchmarks tracked on Vector Wire, Claude Opus 4.6 leads on 43 and Gemini 3 Flash Preview leads on 9 — 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 Gemini 3 Flash Preview

AnthropicvsGoogle53 shared benchmarks43–9 head-to-head

Benchmark	Claude Opus 4.6	Gemini 3 Flash Preview
AA Agentic Index	67.6	49.7
AA Intelligence	44.9	38.7
AA-LCR	74.3	73
AA-Omniscience	13.7	10.1
AIME 2025	99.8	99.7
aime_2026	96.7	96.7
apexAgents	33	27.7
arc_agi_1	93	84.7
ARC-AGI-2	68.8	33.6
arena_elo	1497	1473
arena_text_factuality	1487	1452
arena_vision	1300	1260
Artificial Analysis Coding Index	48.1	42.6
charxiv_rq	69.1	80.3
coding_arena_elo	1545	1387
critpt	12.6	8.6
frontiermath_tier_4	22.9	4.2
gdpval	55.9	35.2
GDPval-AA (Elo)	1619	1204
GPQA Diamond	91.3	90.4
HLE	53.1	43.5
hmmt_feb_2026	96.2	89.4
hmmt_nov_2025	96.3	93.3
IFBench	62.5	78
Legal Agent Benchmark	4.2	0
livebench	76.3	72.4
matharena_visual_math_overall	72.3	85.8
MCP Atlas	76.8	62
mmmlu	91.1	91.8
MMMU-Pro	77.3	81.2
mrcr_v2_8needle_128k_average	84	67.2
OmniDocBench 1.5	86.6	12.1
OmniScience Accuracy	47	53.4
OmniScience Non-Hallucination	37.2	7.6
OSWorld-Verified	72.7	65.1
scicode	52	50.6
screenspot_pro_no_tools	57.7	69.1
simplebench	67.6	61.1
simpleqa_verified	46.5	67.4
swe_bench_bash	75.6	75.8
SWE-bench Multilingual	77.8	72.7
SWE-bench Pro	57.3	49.6
SWE-bench Verified	80.8	78
Terminal-Bench 2.0	65.4	47.6
Terminal-Bench Hard	65.4	38.6
toolathlon	56.8	49.4
vectara_answer_rate	99.8	99.8
vectara_avg_summary_length	137.6	90.2
vectara_factual_consistency	87.8	86.5
vectara_hallucination_rate ↓	12.2	13.5
Vending-Bench 2	801759	363500
τ²-Bench Telecom (AA run)	99.3	80.4
τ³-Bench	72.4	17.5

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.