Zhipu AIGLM 5.2
GLM 5.2 is a Zhipu AI open-weights reasoning model, released 13 Jun 2026. Its best configuration (max reasoning effort) ranks #74 of 756 on the BenchLeader Index at 62.7 ±3.6, in the upper half. The ± is the point: 151 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in instruction following (72) and lowest in knowledge (54). At $2.15 per million tokens blended it is pricier than most ranked models. Output speed of 84 tokens per second puts it faster than most, with a first answer in 26.9 s. It has been measured at 9 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 2 Oct 2026.
- Blended price
- $2.15/M
- $1.40 in · $4.40 out
- Output speed
- 84 tok/s
- measured by Artificial Analysis
- First answer
- 27 s
- first token 0.89 s
- Context
- 1M
- Full answer
- 33 s
- median, reasoning included
- Cost per run
- $1.47
- one full Intelligence Index run
- Released
- 13 Jun 2026
- Overall index63
- Reasoning70
- Coding63
- Agents & tools63
- Maths58
- Knowledge54
- Instruction following72
- Human preference67
- Long context65
- Composite68
Versions
Zhipu AI has shipped 8 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.
Reasoning-effort configurations
The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Not stated” collects results from publishers that did not say which setting they used; for a reasoning model that is usually its thinking mode, but we do not assume it. Pick a setting here or at the top of the page to see its price, speed and category scores.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| no reasoning | 48.4 | #411 | 138 tok/s | 1.31 s | $0.0019 | Agents & tools 45 · Coding 36 · Composite 55 · Knowledge 59 · Long context 46 · Maths 45 · Reasoning 50 |
| minimal | – | – | 72 tok/s | 0.89 s | $0.0019 | |
| low | – | – | 72 tok/s | 0.89 s | $0.0019 | Maths 57 · Reasoning 64 |
| medium | – | – | 72 tok/s | 0.89 s | $0.0019 | |
| high | – | – | 72 tok/s | 0.89 s | $0.0019 | Coding 62 |
| xhigh | – | – | 72 tok/s | 0.89 s | $0.0019 | |
| maxbest | 62.7 | #74 | 84 tok/s | 27 s | $0.0019 | Agents & tools 63 · Coding 63 · Composite 68 · Human preference 67 · Instruction following 72 · Knowledge 54 · Long context 65 · Maths 58 · Reasoning 70 |
| thinking | – | – | 72 tok/s | 0.89 s | $0.0019 | Reasoning 55 |
| not stated | 51.3 | #331 | 72 tok/s | 0.89 s | $0.0019 | Agents & tools 55 · Coding 49 · Composite 44 · Human preference 54 · Knowledge 59 · Maths 47 · Reasoning 53 |
How its index has moved
8 Sept 2026 to 5 Oct 2026- max
- not stated
- no reasoning
The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
Benchmark results
One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | no reasoning | minimal | low | medium | high | xhigh | max | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|---|---|
| GPQA Diamond | 71.2%#161 | – | 87.9%#59 | – | – | – | 91.9%#27 | – | – | Epoch AI Benchmarking Hub | |
| SimpleBench | – | – | – | – | – | – | – | – | 58.8%#32 | SimpleBench | |
| LMArena Hard Prompts | – | – | – | – | – | – | 1496#40 | – | – | LMArena | |
| LiveBench Reasoningnot in index | – | – | – | – | – | – | – | – | 78.6%#51 | LiveBench | |
| GPQA Diamond (AA)not in index | 68.6%#304 | – | – | – | – | – | 89.5%#73 | – | – | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 9.8%#308 | – | – | – | – | – | 41.1%#71 | – | – | Artificial Analysis | |
| GPQA Diamond (Vals)not in index | – | – | – | – | – | – | – | – | 85.6%#44 | Vals AI | |
| Kagi LLM Benchmark | – | – | – | – | – | – | – | 60.0%#50 | 62.6%#44 | Kagi LLM Benchmark | |
| ARC-AGI-1 | – | – | – | – | – | – | – | – | 77.0%#107 | ARC Prize | |
| ARC-AGI-2 | – | – | – | – | – | – | – | – | 22.8%#119 | ARC Prize | |
| CritPt | 3.1%#157 | – | – | – | – | – | 20.9%#51 | – | – | Artificial Analysis | |
| Chess Puzzlesnot in index | 6.0%#150 | – | 14.0%#110 | – | – | – | 21.0%#79 | – | – | Epoch AI Benchmarking Hub | |
| Mystery Game Puzzlesnot in index | 18.0%#64 | 18.0%#64 | 19.0%#59 | 15.0%#85 | – | – | – | – | – | Epoch AI Benchmarking Hub | |
| LMCAnot in index | – | – | – | – | – | – | 45.8%#70 | – | – | Epoch AI Benchmarking Hub | |
| DTBenchnot in index | – | – | – | – | – | – | 93.6%#49 | – | – | Epoch AI Benchmarking Hub |
Coding
| Benchmark | no reasoning | minimal | low | medium | high | xhigh | max | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|---|---|
| SWE-bench Verified (Epoch) | – | – | – | – | – | – | 78.7%#5 | – | – | Epoch AI Benchmarking Hub | |
| SciCode | 36.1%#172 | – | – | – | – | – | 50.5%#94 | – | – | SciCode | |
| WeirdML | – | – | – | – | 67.3%#36 | – | 70.1%#33 | – | – | WeirdML | |
| FrontierCode | 24.5%#33 | – | – | – | – | – | – | – | – | Cognition | |
| LMArena Coding | – | – | – | – | – | – | 1512#51 | – | – | LMArena | |
| LMArena WebDev | – | – | – | – | – | – | 1605#25 | – | – | LMArena | |
| LiveBench Codingnot in index | – | – | – | – | – | – | – | – | 79.7%#22 | LiveBench | |
| SciCode (AA)not in index | – | – | – | – | – | – | 51.2%#89 | – | – | Artificial Analysis | |
| LiveCodeBench | – | – | – | – | – | – | – | – | 69.5%#89 | Vals AI | |
| SWE-bench (Vals)not in index | – | – | – | – | – | – | 82.8%#18 | – | – | Vals AI | |
| Code Migrationnot in index | – | – | – | – | – | – | 37.9%#28 | – | – | Vals AI | |
| ProgramBenchnot in index | – | – | – | – | – | – | 0.5%#13 | – | – | Vals AI | |
| Vibe Code Bench v1.1not in index | – | – | – | – | – | – | 64.0%#40 | – | – | Vals AI | |
| SWE Atlas: Codebase QnAnot in index | – | – | – | – | – | – | – | – | 48.1%#5 | Scale AI SEAL | |
| SWE Atlas: Refactoringnot in index | – | – | – | – | – | – | – | – | 42.4%#9 | Scale AI SEAL | |
| SWE Atlas: Test Writingnot in index | – | – | – | – | – | – | – | – | 41.5%#10 | Scale AI SEAL | |
| Surface Evolver Benchnot in index | – | – | – | – | 55.6%#15 | – | – | – | – | Epoch AI Benchmarking Hub | |
| DeepSWEnot in index | – | – | – | – | 36.3%#59 | – | 43.8%#54 | – | – | Epoch AI Benchmarking Hub | |
| ALE-Benchnot in index | – | – | – | – | 1047.0#47 | – | 1010.2#49 | – | – | Epoch AI Benchmarking Hub | |
| Terminal-Bench 4.0 (AA)not in index | – | – | – | – | – | – | 1.0%#112 | – | – | Artificial Analysis | |
| Terminal-Bench 2.1 (AA)not in index | 51.7%#111 | – | – | – | – | – | 77.9%#54 | – | – | Artificial Analysis |
Agents & tools
Maths
| Benchmark | no reasoning | minimal | low | medium | high | xhigh | max | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|---|---|
| FrontierMath Tiers 1–3 | 42.5%#53 | – | 54.7%#43 | – | – | – | 59.2%#35 | – | – | Epoch AI Benchmarking Hub | |
| FrontierMath Tier 4 | – | – | – | – | – | – | 29.3%#35 | – | – | Epoch AI Benchmarking Hub | |
| OTIS Mock AIME | 28.9%#200 | – | 75.6%#118 | – | – | – | 86.4%#85 | – | – | Epoch AI Benchmarking Hub | |
| ProofBench | – | – | – | – | – | – | 35.0%#39 | – | – | Vals AI | |
| LiveBench Mathematicsnot in index | – | – | – | – | – | – | – | – | 89.8%#34 | LiveBench | |
| AIME 2026 | – | – | – | – | – | – | – | – | 90.0%#27 | MathArena | |
| HMMT February 2026 | – | – | – | – | – | – | – | – | 92.4%#13 | MathArena | |
| MathArena Apex | – | – | – | – | – | – | – | – | 19.8%#15 | MathArena | |
| LMArena Maths | – | – | – | – | – | – | 1488#28 | – | – | LMArena |
Knowledge
| Benchmark | no reasoning | minimal | low | medium | high | xhigh | max | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|---|---|
| SimpleQA Verified | – | – | – | – | – | – | 34.2%#51 | – | – | Epoch AI Benchmarking Hub | |
| LiveBench Data Analysisnot in index | – | – | – | – | – | – | – | – | 73.7%#39 | LiveBench | |
| AA-Omniscience | -6.6#154 | – | – | – | – | – | 4.4#102 | – | – | Artificial Analysis | |
| MMLU-Pro | – | – | – | – | – | – | – | – | 86.7%#38 | Vals AI | |
| LegalBench | – | – | – | – | – | – | – | – | 84.1%#38 | Vals AI | |
| CorpFin | – | – | – | – | – | – | – | – | 66.1%#21 | Vals AI | |
| TaxEval | – | – | – | – | – | – | – | – | 73.3%#48 | Vals AI | |
| Excel Modeling Benchmarknot in index | – | – | – | – | – | – | – | – | 61.5%#30 | Vals AI | |
| MedCodenot in index | – | – | – | – | – | – | – | – | 40.8%#56 | Vals AI | |
| MedScribenot in index | – | – | – | – | – | – | – | – | 83.5%#42 | Vals AI | |
| DrugDiscoveryBenchnot in index | – | – | – | – | – | 36.2%#11 | – | – | – | Scale AI SEAL | |
| EBR-benchnot in index | – | – | – | – | – | – | 9.5%#19 | – | – | Epoch AI Benchmarking Hub | |
| AA-Omniscience: accuracynot in index | 20.3%#282 | – | – | – | – | – | 24.3%#226 | – | – | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 66.3%#56 | – | – | – | – | – | 73.7%#28 | – | – | Artificial Analysis |
Instruction following
| Benchmark | no reasoning | minimal | low | medium | high | xhigh | max | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|---|---|
| LiveBench Languagenot in index | – | – | – | – | – | – | – | – | 76.2%#44 | LiveBench | |
| LiveBench Instruction Followingnot in index | – | – | – | – | – | – | – | – | 62.3%#52 | LiveBench | |
| IFBench | – | – | – | – | – | – | 73.3%#41 | – | – | Artificial Analysis | |
| LMArena Instruction Followingnot in index | – | – | – | – | – | – | 1470#37 | – | – | LMArena |
Human preference
| Benchmark | no reasoning | minimal | low | medium | high | xhigh | max | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|---|---|
| LMArena Text | – | – | – | – | – | – | 1476#32 | – | – | LMArena | |
| EQ-Bench 4 | – | – | – | – | – | – | – | – | 1222#13 | EQ-Bench | |
| LMArena Creative Writingnot in index | – | – | – | – | – | – | 1454#32 | – | – | LMArena | |
| LMArena Multi-turnnot in index | – | – | – | – | – | – | 1475#47 | – | – | LMArena | |
| LMArena Longer Queriesnot in index | – | – | – | – | – | – | 1487#30 | – | – | LMArena |
Long context
| Benchmark | no reasoning | minimal | low | medium | high | xhigh | max | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|---|---|
| AA-LCR | 42.3%#327 | – | – | – | – | – | 78.3%#103 | – | – | Artificial Analysis |
Composite
| Benchmark | no reasoning | minimal | low | medium | high | xhigh | max | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in index | – | – | – | – | – | – | – | – | 151.8#44 | Epoch AI Benchmarking Hub | |
| LiveBench | – | – | – | – | – | – | – | – | 73.2%#45 | LiveBench | |
| AA Intelligence Index | 22.4#186 | – | – | – | – | – | 33.7#90 | – | – | Artificial Analysis |
Where it wins
Benchmarks where this configuration ranks in the top five of every configuration measured.
Where to run it
Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.
| Provider | Speed | First token | Input $/M | Output $/M | Context | Quantisation |
|---|---|---|---|---|---|---|
| Decart Fast | 264 tok/s | 0.59 s | $2.25 | $8.00 | 1.0M | fp4 |
| Baseten Fast | 165 tok/s | 0.64 s | $2.10 | $6.60 | 1.0M | fp8 |
| Baseten Fast | 156 tok/s | 0.61 s | $2.10 | $6.60 | 1.0M | fp8 |
| Parasail | 142 tok/s | 1.61 s | $1.40 | $4.40 | 262k | fp4 |
| CoreWeave | 126 tok/s | 0.52 s | $0.760 | $2.42 | 1.0M | fp4 |
| Mistral (nvfp4) | 114 tok/s | 0.89 s | $1.40 | $4.40 | 1.0M | nvfp4 |
| Wafer | 105 tok/s | 0.52 s | $0.025 | $7.00 | 1.0M | – |
| Mistral | 105 tok/s | 0.51 s | $1.54 | $4.84 | 1.0M | – |
| inference.net | 92 tok/s | 0.68 s | $0.027 | $7.50 | 1.0M | – |
| Alibaba Fast | 90 tok/s | 0.84 s | $2.31 | $7.26 | 1M | fp8 |
| Friendli | 86 tok/s | 0.45 s | $1.40 | $4.40 | 1.0M | – |
| Relace | 81 tok/s | 1.21 s | $0.019 | $16.00 | 1.0M | – |
| Baidu Qianfan (fp4) | 81 tok/s | 1.09 s | $2.25 | $7.88 | 1.0M | fp4 |
| Together | 79 tok/s | 0.43 s | $1.40 | $4.40 | 1.0M | – |
| Phala | 77 tok/s | 1.15 s | $1.26 | $3.00 | 1.0M | fp8 |
| Baseten | 75 tok/s | 2.34 s | $1.40 | $4.40 | 1.0M | fp8 |
| Baidu Qianfan (fp8) | 74 tok/s | 1.18 s | $1.40 | $4.40 | 1.0M | fp8 |
| DeepInfra | 72 tok/s | 1.78 s | $0.563 | $1.80 | 1.0M | fp4 |
| Baseten | 69 tok/s | 2.79 s | $1.40 | $4.40 | 1.0M | fp8 |
| SiliconFlow | 68 tok/s | 1.23 s | $0.700 | $2.20 | 1.0M | fp8 |
| NovitaAI | 64 tok/s | 1.61 s | $0.650 | $2.04 | 1.0M | fp8 |
| Inceptron | 63 tok/s | 0.48 s | $1.39 | $4.39 | 1.0M | fp4 |
| Decart | 62 tok/s | 0.98 s | $0.332 | $2.04 | 1.0M | mxfp4 |
| DigitalOcean | 61 tok/s | 0.82 s | $0.700 | $2.20 | 262k | – |
| Venice | 58 tok/s | 2.20 s | $1.40 | $4.40 | 1M | fp8 |
| Alibaba Cloud Int. | 57 tok/s | 1.57 s | $0.966 | $3.04 | 1M | fp8 |
| Z.ai | 55 tok/s | 3.21 s | $1.40 | $4.40 | 1.0M | fp8 |
| StreamLake | 53 tok/s | 1.36 s | $0.643 | $2.02 | 1.0M | fp8 |
| AtlasCloud | 53 tok/s | 2.17 s | $0.938 | $2.95 | 1.0M | fp8 |
| GMICloud | 48 tok/s | 5.02 s | $1.40 | $4.40 | 1.0M | fp8 |
| Cloudflare | 40 tok/s | 1.67 s | $1.18 | $4.40 | 262k | – |
| Morph | 28 tok/s | 1.00 s | $0.170 | $3.18 | 1.0M | fp8 |
Price history
Listed price per 1M tokens over time, as recorded by OpenRouter for the provider with the longest history.
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $0.260 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0019 | $0.0015 | 30.4 s |
| Summarise a 30-page report | 12,000 / 600 | $0.019 | $0.0092 | 34.0 s |
| Code edit | 6,000 / 1,500 | $0.015 | $0.0099 | 44.7 s |
| Agentic coding session | 60,000 / 4,000 | $0.102 | $0.050 | 1.2 min |
| Structured extraction | 2,000 / 200 | $0.0037 | $0.0020 | 29.2 s |
See also
Data as of 5 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “GLM 5.2: benchmarks, pricing, speed and rank”, https://www.benchleader.com/models/glm-5-2, data as of 5 Oct 2026.