Which is better, GLM-5.2 Full Open Source or Kimi K2.6?

Across 48 shared benchmarks tracked on Vector Wire, GLM-5.2 Full Open Source leads on 35 and Kimi K2.6 leads on 12 — GLM-5.2 Full Open Source leads on more. Strengths differ by axis; the table below shows every head-to-head number with provenance.

SEEDRefresh Models —Deals —Regulatory —Sources —

GLM-5.2 Full Open Source vs Kimi K2.6

Z.aivsMoonshot48 shared benchmarks35–12 head-to-head

Benchmark	GLM-5.2 Full Open Source	Kimi K2.6
AA Agentic Index	45.7	58.7
AA Intelligence	52.6	45.1
AA-LCR	76.7	76.7
AA-Omniscience	4.4	6.4
aime_2026	99.2	96.4
apex_agents	35.6	27.9
apexAgents	33.7	28.5
arena_elo	1471	1461
arena_text_factuality	1460	1455
Artificial Analysis Coding Index	68.8	61.8
coding_arena_elo	1582	1509
critpt	20.9	8
FORTRESS (Adversarial)	71.3	65.6
FORTRESS (Benign)	90	97.2
gdpval	50.3	41.4
GDPval-AA v2	1514	1190
Global-MMLU-Lite	89.2	88.4
GPQA Diamond	91.2	91.1
HiL-Bench	43.7	18.7
HLE	54.7	37.5
HLE (with tools)	54.7	54
hmmt_feb_2026	92.5	94.7
IFBench	73.3	76
imo_answer_bench	91	86
itbenchSre	42.7	31.2
livebench_agentic_coding	51.8	46.9
livebench_coding	79.7	78.6
livebench_data_analysis	73.7	65.1
livebench_instruction_following	62.3	64.4
livebench_language	76.2	75.1
livebench_math	89.8	84.3
livebench_reasoning	78.6	79.4
MCP Atlas	82.6	68.1
OmniScience Accuracy	24.3	32.8
OmniScience Non-Hallucination	73.7	60.7
scicode	50.5	53.5
simpleqa_verified	38.1	38.7
strongreject	98.5	99.8
SWE-bench Pro	62.1	58.6
SWEBench Pro (Public)	62.1	58.6
Tau 3 Banking	26.8	20.6
TauBench V3 - Banking	34.6	23.3
Terminal Bench 2.1 (Best Harness)	82.7	71.3
Terminal-Bench 2.1	82.7	65.9
Terminal-Bench Hard	50.8	43.9
toolathlon	48.2	50
τ²-Bench Telecom (AA run)	99.1	95.9
τ³-Bench	26.8	20.6

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.