AI Models

Standard coding benchmarks

Terminal-Bench 2.1, SWE-bench Pro, NL2Repo, DeepSWE, ProgramBench.

← Back to GLM-5.2 Benchmarks

About this data

On standard coding suites GLM-5.2 is the strongest open-source model and improves sharply on GLM-5.1 (e.g. 81.0 vs 63.5 on Terminal-Bench 2.1). It lands within a few points of Claude Opus 4.8 on Terminal-Bench while staying ahead of Gemini 3.1 Pro, though closed models still lead on several suites.

Standard coding benchmarks

Score (higher = better). GLM-5.2 closes much of the gap to the closed-source frontier.

View data & sources →

Data table

Standard coding benchmarks — coding_std data table (GLM-5.2 Benchmarks)
glm51 glm52 gpt55 gemini opus48 series benchmark source_ref value_basis
63.5 81 84 74 85 coding_std Terminal-Bench 2.1 zai-glm52 Full Benchmark Table — Terminal-Bench 2.1 (Terminus-2)
58.4 62.1 58.6 54.2 69.2 coding_std SWE-bench Pro zai-glm52 Full Benchmark Table — SWE-bench Pro
42.7 48.9 50.7 33.4 69.7 coding_std NL2Repo zai-glm52 Full Benchmark Table — NL2Repo
18 46.2 70 10 58 coding_std DeepSWE zai-glm52 Full Benchmark Table — DeepSWE
50.9 63.7 70.8 39.5 71.9 coding_std ProgramBench zai-glm52 Full Benchmark Table — ProgramBench

Methodology & sources

Last updated: Jul 17, 2026

Methodology

Source-backed values for all six charts come from the Z.ai GLM-5.2 launch post (2026-06-16): long-horizon coding (FrontierSWE, PostTrainBench, SWE-Marathon), standard coding (Terminal-Bench 2.1, SWE-bench Pro, NL2Repo, DeepSWE, ProgramBench), reasoning (HLE, CritPt, AIME, HMMT, GPQA-Diamond, IMOAnswerBench), agentic (MCP-Atlas, Tool-Decathlon), the GLM-5.2 vs GLM-5.1 generational leap, and the MTP acceptance-length ablation. Every numeric point carries a sources[].ref and a value_basis naming the table row. Scores are reproduced verbatim from the Full Benchmark Table; cells the vendor published as "-" (not reported) are omitted, never estimated. CAVEAT: these are the model developer’s self-reported figures under their stated harnesses and prompts (see the post’s footnotes); some competitor HLE/CritPt cells are full-set scores. They are comparison-as-published, not an independent eval. Architecture context (not charted): 1M-token context (up from 200K), IndexShare cuts per-token indexer FLOPs ~2.9× at 1M, MIT-licensed open weights. Re-verified 2026-06-22.

Sources

Comparisons are informative, not definitive. See each source for definitions and limits.

Continue researching

How relevant was this information?