GLM-5 vs GPT-6 Astra
GLM-5 
GPT-6 Astra Verdict
- GPT-6 Astra (max) leads on quality: 70.9 vs 59.1.
- GLM-5 is stronger in human preference, instruction following, long context.
- GPT-6 Astra (max) is stronger in agents & tools, coding, composite, knowledge, maths, reasoning.
- GLM-5 is 13× cheaper ($1.55 vs $20.00 per 1M blended).
- GLM-5 streams 2.4× faster (66 vs 27 tokens per second).
| Metric | GLM-5 | GPT-6 Astra (max) |
|---|---|---|
| BenchLeader Index | 59.1 | 70.9 |
| Agents & tools score | 56.8 | 68.2 |
| Coding score | 52.9 | 78.6 |
| Composite score | 64.3 | 72.9 |
| Human preference score | 65.6 | – |
| Instruction following score | 70.5 | – |
| Knowledge score | 64.3 | 80.0 |
| Long context score | 64.3 | – |
| Maths score | 53.1 | 77.2 |
| Reasoning score | 51.5 | 73.6 |
| Blended price $/M | $1.55 | $20.00 |
| Output speed | 66 tok/s | 27 tok/s |
| Time to first answer | 48.7 s | 337.3 s |
| Context window | 205k | 1.1M |
| GPQA Diamond | 87.8% | 95.8% |
| FrontierMath Tiers 1–3 | – | 93.7% |
| FrontierMath Tier 4 | – | 97.6% |
| OTIS Mock AIME | 80.0% | 100.0% |
| SWE-bench Verified (Epoch) | 72.1% | – |
| SimpleQA Verified | – | 75.6% |
| Terminal-Bench | 52.4% | 58.2% |
| SimpleBench | 53.2% | – |
| SciCode | – | 56.5% |
| WeirdML | 48.2% | – |
| APEX-Agents | 17.2% | – |
| FrontierCode | – | 53.3% |
| Epoch Capabilities Index | 145.9 | – |
| LMArena Text | 1458 | – |
| LMArena Hard Prompts | 1478 | – |
| LMArena Coding | 1498 | – |
| LMArena WebDev | 1436 | 1796 |
| LMArena Agent | – | 12.5 |
| LiveBench | – | 82.2% |
| LiveBench Reasoning | – | 92.7% |
| LiveBench Coding | – | 80.4% |
| LiveBench Agentic Coding | – | 57.3% |
| LiveBench Mathematics | – | 96.8% |
| LiveBench Data Analysis | – | 83.0% |
| LiveBench Language | – | 89.4% |
| AA Intelligence Index | 27.9 | – |
| IFBench | 72.3% | – |
| AA-LCR | 75.7% | – |
| AA-Omniscience | 0.3 | – |
| Terminal-Bench Hard | 43.2% | – |
| GPQA Diamond (AA) | 82.0% | – |
| Humanity's Last Exam (AA) | 29.3% | – |
| τ²-Bench Telecom (AA) | 98.3% | – |
| Terminal-Bench 2.1 (Vals) | – | 87.3% |
| Vals Index | – | 66.6 |
| AIME 2026 | 96.7% | – |
| HMMT February 2026 | 86.4% | – |
| MathArena Apex | 10.9% | – |
| Kagi LLM Benchmark | 51.7% | – |
| ARC-AGI-1 | 44.7% | 97.5% |
| ARC-AGI-2 | 4.9% | 95.0% |
| ARC-AGI-3 | – | 98.6% |
Data as of 2026-09-10. Best configuration of each model; every score links to its source on the model pages.
GLM-5 vs GPT-6 Astra: questions
- Is GLM-5 better than GPT-6 Astra?
- GPT-6 Astra (max) leads on quality: 70.9 vs 59.1. The BenchLeader Index combines every independent quality benchmark; GPT-6 Astra (max) is ahead overall as of 2026-09-10, but check the category scores for your use.
- Is GLM-5 better than GPT-6 Astra for coding?
- GPT-6 Astra scores higher in coding (79 vs 53 on the category index, where 50 is average).
- Is GLM-5 better than GPT-6 Astra for agentic tasks?
- GPT-6 Astra scores higher in agentic tasks (68 vs 57 on the category index, where 50 is average).
- Which is cheaper, GLM-5 or GPT-6 Astra?
- GLM-5 is cheaper: $1.55 against $20.00 per million tokens, blended at three input tokens per output token.
- Which is faster, GLM-5 or GPT-6 Astra?
- GLM-5 streams faster: 66 against 27 output tokens per second.
- Which has the larger context window?
- GPT-6 Astra accepts more context: 1.1M against 205k tokens.