Capability
68.2/100
field median 56.3
#26 of 232 ranked models
Keep up with the models you depend on. Follow price changes, retirements, and API updates.Follow the models you depend on.
Follow model changesReleased Jun 16, 2026 — see all recent releases
Data as of September 14, 2026 · How the score is built
Share or export
Coding ranks #19. Particularly well-suited for software development and code generation tasks.
25 published rows leave some tracked benchmark slots empty. Knowledge is its lowest eligible category at #35.
Each value carries a field reference instead of floating alone. Markers compare this model with the current ranked and priced catalog; they are not absolute quality thresholds.
Capability
68.2/100
field median 56.3
#26 of 232 ranked models
Price
$1.40input / $4.40 output
input median $0.97
blended $2.90
Speed
78tok/s
field median 91.5 tok/s
First token 33.51 s
Context
1Mtokens
field median 256,000
Reported for this model; direct source link not stored
Each axis shows percentile within that category’s eligible cohort. The comparison outline is the median of the six nearest public-score peers; a collapsed vertex means the category is not rank-eligible.
GLM-5.2 category percentile values
The dashed outline is median of 6 nearest peers.
Published API price against the public score. The x-axis uses a log scale; the dashed path marks models that are not beaten by a cheaper, higher-scoring option. Price uses average of published input and output rates.
The chart opens on the current model. Scroll horizontally to inspect the full price axis.
Horizontal: blended price per million tokens, log scale · Vertical: public score
Coverage is split by category so a strong number never hides a thin evidence base. Verified means the row is tied to a published source; provisional rows remain visible but separate.
Each documented value carries its source. Missing fields stay visible as not sourced or not published, rather than disappearing from the page.
Self-host and provider-specific paths stay separate from benchmark evidence so operating constraints are visible before a score becomes the whole decision.
Published weights are available, but BenchLM does not yet have a sourced parameter and VRAM profile for this exact model. Hardware cost estimates stay unavailable until that sizing record is complete.
Estimate VRAM from known parametersScores and ranks appear only where published evidence can be displayed. The table keeps the score, weight, cohort, and evidence state together.
| Category | Score | Rank | Percentile | Weight | Benchmarks | Evidence |
|---|---|---|---|---|---|---|
| AgenticRank #29 of 153Percentile 82ndWeight 22%6 benchmarksVerified | 58.5 | #29 of 153 | 82nd | 22% | 6 benchmarks | Verified |
| CodingRank #19 of 152Percentile 88thWeight 20%8 benchmarksVerified | 61.0 | #19 of 152 | 88th | 20% | 8 benchmarks | Verified |
| ReasoningRank Not rankedWeight 17%1 benchmarkVerified | 74.8 | Not ranked | Not available | 17% | 1 benchmark | Verified |
| KnowledgeRank #35 of 183Percentile 81stWeight 12%6 benchmarksVerified | 60.7 | #35 of 183 | 81st | 12% | 6 benchmarks | Verified |
| MathRank Not rankedWeight 5%4 benchmarksVerified | 80.7 | Not ranked | Not available | 5% | 4 benchmarks | Verified |
| MultilingualWeight 7%0 benchmarksNot measured | Not measured | Not ranked | Not available | 7% | 0 benchmarks | Not measured |
| MultimodalWeight 12%0 benchmarksNot measured | Not measured | Not ranked | Not available | 12% | 0 benchmarks | Not measured |
| Inst. FollowingWeight 5%0 benchmarksNot measured | Not measured | Not ranked | Not available | 5% | 0 benchmarks | Not measured |
Coding opens by default. The marker compares each value with the best source-verified result in the catalog; provisional leaders do not set the reference. Expand the remaining categories for every published row.
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| SWE-bench Pro | Score62.1% | Versus best verified row Best verified: Claude Fable 5.1 · 81.2% | Gap19.1 behind | WeightWeighted 25% | Provider exact |
| LiveCodeBench (Vals)LiveCodeBench, Vals AI run | Score69.5% | Versus best verified row Best verified: Claude Fable 5.1 · 90.5% | Gap21 behind | WeightWeighted 15% | |
| CursorBench 3.2 | Score55.0% | Versus best verified row Best verified: Claude Fable 5.1 · 73.4% | Gap18.4 behind | WeightWeighted 10% | Benchmark exact |
| NL2Repo | Score48.9% | Versus best verified row Best verified: DeepSeek V4.1 Flash · 65.4% | Gap16.5 behind | WeightDisplay only | Provider exact |
| Terminal-Bench 2.0 | Score81.0% | Versus best verified row Best verified: GPT-5.6 Sol · 91.9% | Gap10.9 behind | WeightDisplay only | Provider exact |
| ProgramBenchProgramBench: Can Language Models Rebuild Programs From Scratch? | Score63.7% | Versus best verified row Best verified: Claude Opus 5 · 93.0% | Gap29.3 behind | WeightDisplay only | Provider exact |
| OpenHarmony BenchOpenHarmony Bench v1.0 | Score58.4% | Versus best verified row Best verified: Qwen3.8 Max · 60.8% | Gap2.4 behind | WeightDisplay only | Benchmark exact |
| SWE-bench (Vals)SWE-bench, Vals AI run | Score82.8% | Versus best verified row Best verified: Claude Opus 5 · 97.0% | Gap14.2 behind | WeightDisplay only | Verified |
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| Terminal-Bench 2.0 | Score81% | Versus best verified row Best verified: GPT-5.6 Sol · 91.9% | Gap10.9 behind | WeightWeighted 30% | Provider exact |
| Terminal-Bench 3.0 | Score4.6% | Versus best verified row Best verified: Claude Opus 5 · 42.7% | Gap38.1 behind | WeightDisplay only | Benchmark exact |
| MCP Atlas | Score76.8% | Versus best verified row Best verified: Muse Spark 1.1 · 88.1% | Gap11.3 behind | WeightDisplay only | Provider exact |
| Toolathlon | Score48.2% | Versus best verified row Best verified: Muse Spark 1.1 · 75.6% | Gap27.4 behind | WeightDisplay only | Provider exact |
| ResearchClawBench | Score20.7% | Versus best verified row Best verified: Claude Opus 4.8 · 21.1% | Gap0.4 behind | WeightDisplay only | Benchmark exact |
| Terminal-Bench 2.1 (Vals)Terminal-Bench 2.1, Vals AI run | Score67.8% | Versus best verified row Best verified: GPT-6 Astra · 87.3% | Gap19.5 behind | WeightDisplay only |
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| CritPtCritical Physics Tasks | Score20.9% | Versus best verified row Best verified: GPT-6 Astra · 31.7% | Gap10.8 behind | WeightDisplay only | Provider exact |
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| HLEHumanity's Last Exam | Score54.7% | Versus best verified row Best verified: Claude Fable 5.1 · 65% | Gap10.3 behind | WeightWeighted 35% | Provider exact |
| HLE w/o toolsHumanity's Last Exam without tools | Score40.5% | Versus best verified row Best verified: Claude Fable 5.1 · 60.9% | Gap20.4 behind | WeightWeighted 10% | Provider exact |
| MMLU-Pro (Vals)MMLU-Pro, Vals AI run | Score86.7% | Versus best verified row Best verified: Claude Fable 5.1 · 92.4% | Gap5.7 behind | WeightWeighted 10% | Verified |
| GPQAGraduate-Level Google-Proof Q&A | Score91.2% | Versus best verified row Best verified: GPT-6 Astra · 96% | Gap4.8 behind | WeightWeighted 7% | Provider exact |
| GPQA-DGPQA Diamond | Score91.2% | Versus best verified row Best verified: GPT-6 Astra · 96.0% | Gap4.8 behind | WeightDisplay only | Provider exact |
| GPQA Diamond (Vals)GPQA Diamond, Vals AI run | Score85.6% | Versus best verified row Best verified: Gemini 3.1 Pro · 95.5% | Gap9.9 behind | WeightDisplay only |
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| AIME26AIME 2026 | Score99.2% | Versus best verified row Best verified: GLM-5.2 · 99.2% | GapBest verified | WeightWeighted 25% | Provider exact |
| HMMT Feb 2026Harvard-MIT Mathematics Tournament February 2026 | Score92.5% | Versus best verified row Best verified: Qwen3.7 Max · 97.1% | Gap4.6 behind | WeightWeighted 25% | Provider exact |
| HMMT Nov 2025Harvard-MIT Mathematics Tournament November 2025 | Score94.4% | Versus best verified row Best verified: Qwen3.6 Plus · 94.6% | Gap0.2 behind | WeightDisplay only | Provider exact |
| MMAnswerBench | Score91.0% | Versus best verified row Best verified: GLM-5.2 · 91.0% | GapBest verified | WeightDisplay only | Provider exact |
The sequence follows explicit supersedes links. A successor's displayed score stays at least 0.1 points above its predecessor; raw benchmark rows do not move. Scores and prices remain blank when the corresponding public row or first-party rate is unavailable.
Mar 1, 2026
GLM-5Score 61.5 · $1 / $3.2
Apr 7, 2026
GLM-5.1Score 63.3 · $1.4 / $4.4
Jun 16, 2026 · you are here
GLM-5.2Score 68.2 · $1.4 / $4.4
Aug 14, 2026
GLM-5.3Score 68.4 · Price not listed
flagship · 5.2
The visual layer above carries the decisions. These notes preserve the model, ranking, coverage, and family context behind the numbers.
GLM-5.2 ranks #26 of 232 on the public leaderboard with a score of 68.19/100. Its source-verified position is #20 of 130.
GLM-5.2 is a open weight model with a 1M context window. It uses an explicit reasoning mode, which can improve complex problem solving while adding latency and token use.
Official exact-value snapshot from Z.AI's GLM-5.2 launch materials and Hugging Face model card. BenchLM maps directly comparable rows into existing keys; DeepSWE, FrontierSWE, PostTrainBench, SWE-Marathon, and other unsupported long-horizon rows remain excluded or external-only until stable local keys exist.
GLM-5.2 sits in the GLM-5 family with GLM-5, GLM-5.1, GLM-5.3, GLM-5.3-Flash, GLM-5-Turbo, GLM-5V-Turbo, GLM-5 (Reasoning). Its explicit predecessor is GLM-5.1. 25 of 435 tracked benchmark slots currently have displayable evidence. Missing categories stay blank.
Its strongest eligible category is Coding at #19, while its lowest eligible position is Knowledge at #35. particularly well-suited for software development and code generation tasks.
Z.AI · Model release
Radar confirmed these at the source. Use GLM-5.2 in your work? Explore Radar to follow supported changes and choose your alerts.
GLM-5.2 ranks #26 out of 232 models on the public BenchAlign leaderboard, with a score of 68.19/100. Its evidence status is Supported, and this profile shows 25 source-displayable benchmark rows. The label describes evidence depth, not a provider quality claim; inspect category rows before choosing a workload.
GLM-5.2 ranks #35 out of 183 eligible models for knowledge and understanding, with a public category score of 60.7/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
GLM-5.2 ranks #19 out of 152 eligible models for coding and programming, with a public category score of 61/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
GLM-5.2 has source-displayable benchmark coverage for mathematics, but the public category table does not assign it a rank there. The individual rows remain available for inspection. A missing category position means the evidence threshold was not met; it does not convert the model's unmeasured work into a zero.
GLM-5.2 has source-displayable benchmark coverage for reasoning and logic, but the public category table does not assign it a rank there. The individual rows remain available for inspection. A missing category position means the evidence threshold was not met; it does not convert the model's unmeasured work into a zero.
GLM-5.2 ranks #29 out of 153 eligible models for agentic tool use and computer tasks, with a public category score of 58.5/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
GLM-5.2 is an open-weight model from Z.AI. Its weights can be downloaded for local or hosted deployment, subject to the published license. Open weight does not automatically mean open source: training data and training code may remain private, and commercial restrictions can still apply.
GLM-5.2 belongs to the GLM-5 family. Related tracked variants include GLM-5, GLM-5.1, GLM-5.3, plus 4 more. A sibling link indicates shared lineage or a documented configuration relationship; it does not mean the variants have identical pricing, context limits, benchmark evidence, or deployment behavior. Compare before switching.
No. GLM-5.2 currently has 42 source-displayable rows across 435 tracked benchmark slots. The profile exposes published, non-generated evidence and leaves missing categories blank until an exact evaluation is available. Coverage describes how much was measured; it is not a penalty added to an individual benchmark result.
GLM-5.2 has a reported context window of 1M in the exact-model catalog record. The value stays visible, but the profile marks its source link as unavailable instead of presenting it as directly documented. Maximum output length remains separate because providers often publish a different limit.
Related resources
Last updated September 14, 2026. Runtime fields remain blank until a sourced snapshot exists.
Get one weekly email when material rank, price, availability, or benchmark evidence changes are worth revisiting.
Read a sample issueJoin 2,000+ readers.