Reasoning benchmarks (e.g. HLE).
Reasoning benchmarks
Takeaway: The widest remaining gap to Opus/Gemini.
Reasoning
HLE, CritPt, AIME 2026, HMMT, GPQA-Diamond, IMOAnswerBench.
Takeaway: The widest remaining gap to Opus/Gemini.
Reasoning
The widest remaining gap to Opus/Gemini.
Rendering the interactive chart…
Expand data table
Caveat: Some competitor scores are marked full-set vs public-set, not always like-for-like.
Unknown values are hatched and never treated as zero. Zero remains a verified value.
Data table
Download CSV| Glm51 | Glm52 | Gpt55 | Gemini | Opus48 | Series | Benchmark | Source | Value basis |
|---|---|---|---|---|---|---|---|---|
| 31 | 40.5 | 41.4 | 45 | 49.8 | reasoning | HLE | Z.ai: GLM-5.2: Built for Long-Horizon Tasks (2026-06-16) | Actual |
| 52.3 | 54.7 | 52.2 | 51.4 | 57.9 | reasoning | HLE w/ Tools | Z.ai: GLM-5.2: Built for Long-Horizon Tasks (2026-06-16) | Actual |
| 4.6 | 20.9 | 27.1 | 17.7 | 20.9 | reasoning | CritPt | Z.ai: GLM-5.2: Built for Long-Horizon Tasks (2026-06-16) | Actual |
| 95.3 | 99.2 | 98.3 | 98.2 | 95.7 | reasoning | AIME 2026 | Z.ai: GLM-5.2: Built for Long-Horizon Tasks (2026-06-16) | Actual |
| 82.6 | 92.5 | 96.7 | 87.3 | 96.7 | reasoning | HMMT Feb 2026 | Z.ai: GLM-5.2: Built for Long-Horizon Tasks (2026-06-16) | Actual |
| 86.2 | 91.2 | 93.6 | 94.3 | 93.6 | reasoning | GPQA-Diamond | Z.ai: GLM-5.2: Built for Long-Horizon Tasks (2026-06-16) | Actual |
| 83.8 | 91 | Not published | 81 | 83.5 | reasoning | IMOAnswerBench | Z.ai: GLM-5.2: Built for Long-Horizon Tasks (2026-06-16) | Actual |
How this evidence was built
What it measures
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.
Key context
- Every numeric point carries a source citation and a basis label naming the table row.
- Scores are reproduced verbatim from the Full Benchmark Table; cells the vendor published as "-" (not reported) are omitted, never estimated.
Key numbers
- FrontierSWE: glm-5.2 74.4 vs Opus 4.8 75.1 vs gpt-5.5 72.6 vs Gemini 39.6 vs glm-5.1 30.5 (zai-glm52).
- Terminal-Bench 2.1: glm-5.2 81 vs Opus 4.8 85 vs gpt-5.5 84 vs Gemini 74 (zai-glm52).
- Generational leap: swe-Marathon 13.0 (5.2) vs 1.0 (5.1) (zai-glm52).
- Mtp ablation quantifies the multi-token-prediction contribution to accepted length (zai-glm52).
Caveats
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.
Caveat: Some competitor scores are marked full-set vs public-set, not always like-for-like.
Do not use this to
- Independent verification: every figure is vendor self-reported (zai-glm52).
- Real-world reliability, latency or cost (benchmark scores only).
Scope
June 2026
Sources
Last verified 10 Sept 2026