Claude Sonnet 5.5 vs GLM-5
Claude Sonnet 5.5 
GLM-5 Verdict
- Claude Sonnet 5.5 (max) leads on quality: 67.6 vs 58.9.
- Claude Sonnet 5.5 (max) is stronger in agents & tools, coding, composite, knowledge, long context, maths, reasoning.
- GLM-5 (thinking) is stronger in instruction following.
- GLM-5 (thinking) is 2.6× cheaper ($1.55 vs $4.00 per 1M blended).
- Claude Sonnet 5.5 (max) streams 2.0× faster (141 vs 70 tokens per second).
| Metric | Claude Sonnet 5.5 (max) | GLM-5 (thinking) |
|---|---|---|
| BenchLeader Index | 67.6 | 58.9 |
| Agents & tools score | 69.5 | 57.9 |
| Coding score | 69.8 | 59.2 |
| Composite score | 72.6 | 61.1 |
| Knowledge score | 65.4 | 58.5 |
| Long context score | 66.6 | 63.0 |
| Maths score | 72.9 | 63.2 |
| Reasoning score | 81.1 | 52.3 |
| Instruction following score | – | 70.7 |
| Blended price $/M | $4.00 | $1.55 |
| Output speed | 141 tok/s | 70 tok/s |
| Time to first answer | 487.1 s | 46.0 s |
| Context window | 1M | 205k |
| GPQA Diamond | 95.6% | – |
| FrontierMath Tiers 1–3 | 88.8% | – |
| FrontierMath Tier 4 | 80.5% | – |
| OTIS Mock AIME | 100.0% | – |
| SimpleQA Verified | 46.5% | – |
| Terminal-Bench | 61.8% | – |
| SciCode | 61.0% | – |
| APEX-Agents | 75.5% | – |
| ProofBench | 100.0% | – |
| LMArena Agent | 12 | – |
| LiveBench | 75.7% | – |
| LiveBench Reasoning | 91.6% | – |
| LiveBench Coding | 91.4% | – |
| LiveBench Agentic Coding | 56.3% | – |
| LiveBench Mathematics | 96.1% | – |
| LiveBench Data Analysis | 59.5% | – |
| LiveBench Language | 78.0% | – |
| LiveBench Instruction Following | 56.8% | – |
| AA Intelligence Index v4.3.2 | 56 | 27.9 |
| IFBench | – | 72.3% |
| AA-LCR | 82.7% | 75.7% |
| AA-Omniscience | 32.3 | 0.3 |
| Terminal-Bench Hard | – | 43.2% |
| GPQA Diamond (AA) | – | 82.0% |
| Humanity's Last Exam (AA) | 55.0% | 29.3% |
| SciCode (AA) | 61.0% | – |
| τ²-Bench Telecom (AA) | – | 98.3% |
| AIME (Vals) | – | 91.7% |
| LiveCodeBench | – | 81.9% |
| MMLU-Pro | – | 86.0% |
| LegalBench | – | 84.1% |
| CorpFin | – | 62.9% |
| TaxEval | – | 70.0% |
| MedQA | – | 94.3% |
| SWE-bench (Vals) | – | 71.4% |
| GPQA Diamond (Vals) | – | 83.3% |
| Kagi LLM Benchmark | – | 75.0% |
| τ²-bench | – | 81.0% |
| CritPt | 31.4% | 2.0% |
| GDPval-AA v2.1 | 67.0% | – |
| Analyst Agent (AA) | 57.5% | – |
| APEX-Agents (AA) | – | 14.4% |
| CaseLaw v2 | – | 52.5% |
| Terminal-Bench 2.0 (Vals) | – | 49.4% |
| Vibe Code Bench v1.1 | – | 23.4% |
| Mystery Game Puzzles | 65.0% | – |
| CursorBench | 55.5% | – |
| FrontierSWE | 61.9% | – |
| Terminal-Bench 4.0 (AA) | 63.6% | – |
| AutomationBench | 71.8% | – |
| GDP.pdf | 25.8% | – |
| MLCR | 75.0% | – |
| Harvey LAB | 2.8% | – |
| AA-Omniscience: accuracy | 54.0% | 26.3% |
| AA-Omniscience: non-hallucination | 53.0% | 64.7% |
| AA-Briefcase v1.1 | 1823 | – |
Data as of 2026-10-11. Best configuration of each model; every score links to its source on the model pages.
Claude Sonnet 5.5 vs GLM-5: questions
- Is Claude Sonnet 5.5 better than GLM-5?
- Claude Sonnet 5.5 (max) leads on quality: 67.6 vs 58.9. The BenchLeader Index combines every independent quality benchmark; Claude Sonnet 5.5 (max) is ahead overall as of 2026-10-11, but check the category scores for your use.
- Is Claude Sonnet 5.5 better than GLM-5 for coding?
- Claude Sonnet 5.5 scores higher in coding (70 vs 59 on the category index, where 50 is average).
- Is Claude Sonnet 5.5 better than GLM-5 for agentic tasks?
- Claude Sonnet 5.5 scores higher in agentic tasks (70 vs 58 on the category index, where 50 is average).
- Which is cheaper, Claude Sonnet 5.5 or GLM-5?
- GLM-5 is cheaper: $1.55 against $4.00 per million tokens, blended at three input tokens per output token.
- Which is faster, Claude Sonnet 5.5 or GLM-5?
- Claude Sonnet 5.5 streams faster: 141 against 70 output tokens per second.
- Which has the larger context window?
- Claude Sonnet 5.5 accepts more context: 1M against 205k tokens.