Claude Opus 4.7 vs GLM-5
Claude Opus 4.7 
GLM-5 Verdict
- Claude Opus 4.7 leads on quality: 65.6 vs 59.1.
- Claude Opus 4.7 is stronger in agents & tools, coding, composite, human preference, knowledge, long context, maths, multimodal, reasoning.
- GLM-5 is stronger in instruction following.
- GLM-5 is 6.5× cheaper ($1.55 vs $10.00 per 1M blended).
- GLM-5 streams 1.5× faster (66 vs 43 tokens per second).
| Metric | Claude Opus 4.7 | GLM-5 |
|---|---|---|
| BenchLeader Index | 65.6 | 59.1 |
| Agents & tools score | 69.4 | 56.8 |
| Coding score | 64.8 | 52.9 |
| Composite score | 80.4 | 64.3 |
| Human preference score | 69.1 | 65.6 |
| Instruction following score | 58.7 | 70.5 |
| Knowledge score | 66.5 | 64.3 |
| Long context score | 65.9 | 64.3 |
| Maths score | 65.4 | 53.1 |
| Multimodal score | 65.7 | – |
| Reasoning score | 66.6 | 51.5 |
| Blended price $/M | $10.00 | $1.55 |
| Output speed | 43 tok/s | 66 tok/s |
| Time to first answer | 21.7 s | 48.7 s |
| Context window | 1M | 205k |
| GPQA Diamond | – | 87.8% |
| OTIS Mock AIME | – | 80.0% |
| SWE-bench Verified (Epoch) | – | 72.1% |
| Humanity's Last Exam | 36.2% | – |
| Terminal-Bench | 80.2% | 52.4% |
| SimpleBench | 61.7% | 53.2% |
| WeirdML | 76.4% | 48.2% |
| APEX-Agents | – | 17.2% |
| FrontierCode | 38.5% | – |
| GSO-Bench | 44.1% | – |
| Epoch Capabilities Index | 156.3 | 145.9 |
| LMArena Text | 1495 | 1458 |
| LMArena Hard Prompts | 1519 | 1478 |
| LMArena Coding | 1547 | 1498 |
| LMArena WebDev | 1557 | 1436 |
| LMArena Vision | 1317 | – |
| AA Intelligence Index | 40.7 | 27.9 |
| IFBench | 58.6% | 72.3% |
| AA-LCR | 78.7% | 75.7% |
| MMMU-Pro | 78.8% | – |
| AA-Omniscience | 27.3 | 0.3 |
| Terminal-Bench Hard | 51.5% | 43.2% |
| GPQA Diamond (AA) | 91.4% | 82.0% |
| Humanity's Last Exam (AA) | 42.3% | 29.3% |
| τ²-Bench Telecom (AA) | 88.6% | 98.3% |
| AIME (Vals) | 96.3% | – |
| LiveCodeBench | 85.1% | – |
| MMLU-Pro | 89.9% | – |
| IOI | 47.1% | – |
| LegalBench | 85.3% | – |
| CorpFin | 66.1% | – |
| TaxEval | 75.3% | – |
| Terminal-Bench 2.1 (Vals) | 68.5% | – |
| SWE-bench (Vals) | 82.0% | – |
| GPQA Diamond (Vals) | 90.2% | – |
| Vals Index | 56.1 | – |
| AIME 2026 | – | 96.7% |
| HMMT February 2026 | – | 86.4% |
| MathArena Apex | – | 10.9% |
| HiL-Bench | 41.7% | – |
| EQ-Bench 4 | 1311 | – |
| Kagi LLM Benchmark | 73.3% | 51.7% |
| ARC-AGI-1 | – | 44.7% |
| ARC-AGI-2 | – | 4.9% |
Data as of 2026-09-10. Best configuration of each model; every score links to its source on the model pages.
Claude Opus 4.7 vs GLM-5: questions
- Is Claude Opus 4.7 better than GLM-5?
- Claude Opus 4.7 leads on quality: 65.6 vs 59.1. The BenchLeader Index combines every independent quality benchmark; Claude Opus 4.7 is ahead overall as of 2026-09-10, but check the category scores for your use.
- Is Claude Opus 4.7 better than GLM-5 for coding?
- Claude Opus 4.7 scores higher in coding (65 vs 53 on the category index, where 50 is average).
- Is Claude Opus 4.7 better than GLM-5 for agentic tasks?
- Claude Opus 4.7 scores higher in agentic tasks (69 vs 57 on the category index, where 50 is average).
- Which is cheaper, Claude Opus 4.7 or GLM-5?
- GLM-5 is cheaper: $1.55 against $10.00 per million tokens, blended at three input tokens per output token.
- Which is faster, Claude Opus 4.7 or GLM-5?
- GLM-5 streams faster: 66 against 43 output tokens per second.
- Which has the larger context window?
- Claude Opus 4.7 accepts more context: 1M against 205k tokens.