Model comparison
Claude Opus 4.8 vs GLM-5.2
Claude Opus 4.8 is the stronger model overall, scoring 60.7 to 51.1 on the Noometry Index. GLM-5.2 costs 4.7× less per token, which makes it the better buy when Claude Opus 4.8's lead doesn't matter for your workload.
Last verified . 50 shared benchmarks.
Summary
- They share 50 benchmarks with published results for both. Claude Opus 4.8 scores higher in 8 categories and GLM-5.2 in 1 category; 6 gaps are clear of the uncertainty.
- The widest gap is in math, where Claude Opus 4.8 leads 78.4 to 55.7.
- The biggest single-benchmark swing is GBAEval: 70.9% for Claude Opus 4.8 and 0% for GLM-5.2.
- GLM-5.2 is cheaper at $1.40 / $4.40 per million input/output tokens, against $5 / $25 for Claude Opus 4.8.
- GLM-5.2 has downloadable open weights; the other is API-only.
Side by side
| Claude Opus 4.8 | GLM-5.2 | |
|---|---|---|
| Provider | Anthropic | Z.ai (Zhipu) |
| Noometry Index | 60.7 | 51.1 |
| Released | 2026-05-28 | 2026-06-13 |
| Weights | Proprietary | Open |
| Context window | 1M | 1M |
| Max output | 128K | 131K |
| Input $ / M tokens | $5 | $1.40 |
| Output $ / M tokens | $25 | $4.40 |
| Results tracked | 65 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Opus 4.8 leads
Claude Opus 4.8: 59.9 (#12), GLM-5.2: 51.3 (#41)
| Benchmark | Claude Opus 4.8 | GLM-5.2 |
|---|---|---|
| DeepSWE | 59% | 43.8% |
| FrontierCode | 46.5% | 24.5% |
| LMArena WebDev | 1556 | 1603 |
| SciCode | 53.5% | 50.5% |
| WeirdML | 82.9% | 70.1% |
| LMArena Coding | 1490 | 1485 |
| ALE-Bench | 1,564 | 1,047 |
| SWE-bench Verified | — | 78.7% |
| GSO | 47.1% | — |
Agentic & Tool Use Claude Opus 4.8 leads
Claude Opus 4.8: 47.6 (#11), GLM-5.2: 32.4 (#63)
| Benchmark | Claude Opus 4.8 | GLM-5.2 |
|---|---|---|
| APEX-Agents | 48.9% | 45.2% |
| τ²-bench Banking | 39.7% | 37.1% |
| PostTrainBench | 33.8% | 31.7% |
| GBAEval | 70.9% | 0% |
| Vending-Bench 2 | 5,787 | 8,314 |
| OSWorld 2.0 | 20.6% | — |
| Remote Labor Index | 8.3% | — |
| DeepResearch Bench | 50.2% | — |
| GDP.pdf | 24% | — |
| LMArena Search | 1204 | — |
Reasoning Claude Opus 4.8 leads
Claude Opus 4.8: 64.7 (#16), GLM-5.2: 42.3 (#52)
| Benchmark | Claude Opus 4.8 | GLM-5.2 |
|---|---|---|
| ARC-AGI-2 | 72.1% | 22.8% |
| SimpleBench | 64.8% | 58.8% |
| Kagi LLM Benchmark | 88.8% | 62.6% |
| NYT Connections (extended) | 91.1% | 74.3% |
| ARC-AGI-1 | 92.5% | 77% |
| CritPt | 20.9% | 20.9% |
| Chess Puzzles | 34% | 21% |
| EBR-Bench | 28.6% | 9.5% |
| LMArena Hard Prompts | 1482 | 1480 |
| Mystery Game Puzzles | 36% | 19% |
| DTBench | 94.9% | 93.6% |
| LMCA | 57.5% | 45.8% |
| Surface Evolver Bench | 87.5% | 55.6% |
| Epoch Capabilities Index | 158.21 | 151.78 |
| EnigmaEval | 23.5% | — |
| Bench to the Future 3 | 0.14 | — |
| ForecastBench | 59.9 | — |
Math Claude Opus 4.8 leads
Claude Opus 4.8: 78.4 (#13), GLM-5.2: 55.7 (#43)
| Benchmark | Claude Opus 4.8 | GLM-5.2 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 80% | 59.2% |
| FrontierMath Tier 4 | 56.1% | 29.3% |
| MathArena Final-Answer Competitions | 91.8% | 67.6% |
| OTIS Mock AIME 2024-2025 | 98.3% | 86.4% |
| ProofBench | 69% | 35% |
| LMArena Math | 1487 | 1482 |
| FrontierMath (Feb 2025 set) | 47.2% | — |
| FrontierMath Tier 4 (v1) | 31.3% | — |
Knowledge Claude Opus 4.8 leads
Claude Opus 4.8: 61.3 (#29), GLM-5.2: 57.1 (#40)
| Benchmark | Claude Opus 4.8 | GLM-5.2 |
|---|---|---|
| GPQA Diamond | 91% | 91.9% |
| SimpleQA Verified | 53% | 34.2% |
| LMArena Expert | 1502 | 1486 |
Multimodal Not comparable
Claude Opus 4.8: 42.9 (#26), GLM-5.2: —
| Benchmark | Claude Opus 4.8 | GLM-5.2 |
|---|---|---|
| LMArena Vision | 1294 | — |
| Blueprint-Bench 2 | 14.5% | — |
| Furniture Assembly | 42.5% | — |
| LMArena Document | 1475 | — |
Multilingual Too close to call
Claude Opus 4.8: 55.2 (#33), GLM-5.2: 55.8 (#26)
| Benchmark | Claude Opus 4.8 | GLM-5.2 |
|---|---|---|
| LMArena Non-English | 1450 | 1459 |
| LMArena Chinese | 1507 | 1519 |
| LMArena French | 1481 | 1479 |
| LMArena German | 1472 | 1468 |
| LMArena Japanese | 1440 | 1451 |
| LMArena Korean | 1432 | 1445 |
| LMArena Russian | 1474 | 1466 |
| LMArena Spanish | 1466 | 1477 |
Instruction Following Too close to call
Claude Opus 4.8: 77.4 (#24), GLM-5.2: 76.9 (#34)
| Benchmark | Claude Opus 4.8 | GLM-5.2 |
|---|---|---|
| LMArena Instruction Following | 1476 | 1465 |
Long Context Too close to call
Claude Opus 4.8: 45.4 (#35), GLM-5.2: 45.3 (#43)
| Benchmark | Claude Opus 4.8 | GLM-5.2 |
|---|---|---|
| LMArena Longer Query | 1483 | 1479 |
Writing & Preference Claude Opus 4.8 leads
Claude Opus 4.8: 72.0 (#16), GLM-5.2: 70.4 (#21)
| Benchmark | Claude Opus 4.8 | GLM-5.2 |
|---|---|---|
| LMArena Text | 1461 | 1470 |
| LMArena Creative Writing | 1454 | 1462 |
| EQ-Bench Creative Writing | 1840 | 1757 |
| EQ-Bench 4 | 1281 | 1222 |
| LMArena Multi-Turn | 1476 | 1469 |
Frequently asked questions
Is Claude Opus 4.8 better than GLM-5.2?
Claude Opus 4.8 is the stronger model overall, scoring 60.7 to 51.1 on the Noometry Index. GLM-5.2 costs 4.7× less per token, which makes it the better buy when Claude Opus 4.8's lead doesn't matter for your workload.
Which is cheaper, Claude Opus 4.8 or GLM-5.2?
GLM-5.2 is cheaper. It lists at $1.40 per million input tokens and $4.40 per million output tokens; Claude Opus 4.8 lists at $5 and $25.
Is Claude Opus 4.8 or GLM-5.2 better for coding?
Claude Opus 4.8 scores higher on coding benchmarks: 59.9 versus 51.3 in the Noometry coding category.
Which has the bigger context window?
Both accept 1M tokens.
How many benchmarks do Claude Opus 4.8 and GLM-5.2 share?
50 benchmarks have published results for both models. Claude Opus 4.8 has 65 scored results on Noometry and GLM-5.2 has 51.