Model comparison
GLM-4.5V vs GPT-6 Astra
GPT-6 Astra is the stronger model overall, scoring 70.8 to 39.8 on the Noometry Index. GLM-4.5V costs 22× less per token, which makes it the better buy when GPT-6 Astra's lead doesn't matter for your workload.
Last verified . 14 shared benchmarks.
Summary
- They share 14 benchmarks with published results for both. GLM-4.5V scores higher in 0 categories and GPT-6 Astra in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-6 Astra leads 85.1 to 27.4.
- GLM-4.5V is cheaper at $0.60 / $1.80 per million input/output tokens, against $10 / $50 for GPT-6 Astra.
- GPT-6 Astra accepts more context: 1.05M tokens versus 64K.
- GLM-4.5V has downloadable open weights; the other is API-only.
Side by side
| GLM-4.5V | GPT-6 Astra | |
|---|---|---|
| Provider | Z.ai (Zhipu) | OpenAI |
| Noometry Index | 39.8 | 70.8 |
| Released | 2025-08-11 | 2026-09-03 |
| Weights | Open | Proprietary |
| Context window | 64K | 1.05M |
| Max output | 16K | 128K |
| Input $ / M tokens | $0.60 | $10 |
| Output $ / M tokens | $1.80 | $50 |
| Results tracked | 15 | 56 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-6 Astra leads
GLM-4.5V: 39.5 (#155), GPT-6 Astra: 73.7 (#2)
| Benchmark | GLM-4.5V | GPT-6 Astra |
|---|---|---|
| LMArena Coding | 1347 | 1487 |
| DeepSWE | — | 74.1% |
| FrontierCode | — | 53.3% |
| LMArena WebDev | — | 1786 |
| FrontierSWE | — | 65.5% |
| SciCode | — | 56.5% |
| GSO | — | 79.4% |
| WeirdML | — | 93.6% |
| MirrorCode | — | 46.7% |
| ALE-Bench | — | 2,951 |
Agentic & Tool Use Not comparable
GLM-4.5V: —, GPT-6 Astra: 52.9 (#3)
| Benchmark | GLM-4.5V | GPT-6 Astra |
|---|---|---|
| APEX-Agents | — | 64.7% |
| Remote Labor Index | — | 20.8% |
| BALROG | — | 68.3% |
| GDP.pdf | — | 34.2% |
| Vending-Bench 2 | — | 15,515 |
Reasoning GPT-6 Astra leads
GLM-4.5V: 27.4 (#119), GPT-6 Astra: 85.1 (#1)
| Benchmark | GLM-4.5V | GPT-6 Astra |
|---|---|---|
| LMArena Hard Prompts | 1334 | 1462 |
| ARC-AGI-2 | — | 95% |
| Kagi LLM Benchmark | 59.8% | — |
| NYT Connections (extended) | — | 98.1% |
| ARC-AGI-1 | — | 98.5% |
| CritPt | — | 31.7% |
| Chess Puzzles | — | 72% |
| EBR-Bench | — | 76.2% |
| Mystery Game Puzzles | — | 84% |
| DTBench | — | 97.3% |
| LMCA | — | 64.4% |
| Bench to the Future 3 | — | 0.14 |
| Epoch Capabilities Index | — | 166.45 |
Math GPT-6 Astra leads
GLM-4.5V: 37.4 (#159), GPT-6 Astra: 93.5 (#2)
| Benchmark | GLM-4.5V | GPT-6 Astra |
|---|---|---|
| LMArena Math | 1354 | 1465 |
| FrontierMath (Tiers 1-3) | — | 93.7% |
| FrontierMath Tier 4 | — | 97.6% |
| OTIS Mock AIME 2024-2025 | — | 100% |
| ProofBench | — | 99% |
| FrontierMath Erdős | — | 2.9% |
Knowledge GPT-6 Astra leads
GLM-4.5V: 37.5 (#156), GPT-6 Astra: 75.3 (#1)
| Benchmark | GLM-4.5V | GPT-6 Astra |
|---|---|---|
| LMArena Expert | 1353 | 1483 |
| GPQA Diamond | — | 95.8% |
| Humanity's Last Exam | — | 54.8% |
| SimpleQA Verified | — | 75.6% |
| Vectara Hallucination Rate | — | 8.7% |
Multimodal GPT-6 Astra leads
GLM-4.5V: 34.3 (#92), GPT-6 Astra: 55.0 (#3)
| Benchmark | GLM-4.5V | GPT-6 Astra |
|---|---|---|
| LMArena Vision | 1154 | 1281 |
| Blueprint-Bench 2 | — | 49.7% |
| Furniture Assembly | — | 80% |
| LMArena Document | — | 1468 |
Multilingual GPT-6 Astra leads
GLM-4.5V: 44.6 (#177), GPT-6 Astra: 53.7 (#61)
| Benchmark | GLM-4.5V | GPT-6 Astra |
|---|---|---|
| LMArena Non-English | 1303 | 1430 |
| LMArena Chinese | 1337 | 1484 |
| LMArena Russian | 1298 | 1436 |
| LMArena Spanish | 1336 | 1407 |
| LMArena French | — | 1456 |
| LMArena German | — | 1440 |
| LMArena Japanese | — | 1379 |
| LMArena Korean | — | 1426 |
Instruction Following GPT-6 Astra leads
GLM-4.5V: 69.2 (#175), GPT-6 Astra: 76.3 (#44)
| Benchmark | GLM-4.5V | GPT-6 Astra |
|---|---|---|
| LMArena Instruction Following | 1311 | 1450 |
Long Context GPT-6 Astra leads
GLM-4.5V: 39.6 (#171), GPT-6 Astra: 44.5 (#62)
| Benchmark | GLM-4.5V | GPT-6 Astra |
|---|---|---|
| LMArena Longer Query | 1304 | 1456 |
Writing & Preference GPT-6 Astra leads
GLM-4.5V: 52.5 (#170), GPT-6 Astra: 75.3 (#7)
| Benchmark | GLM-4.5V | GPT-6 Astra |
|---|---|---|
| LMArena Text | 1333 | 1441 |
| LMArena Creative Writing | 1295 | 1418 |
| LMArena Multi-Turn | 1332 | 1448 |
| EQ-Bench Creative Writing | — | 2173 |
Frequently asked questions
Is GLM-4.5V better than GPT-6 Astra?
GPT-6 Astra is the stronger model overall, scoring 70.8 to 39.8 on the Noometry Index. GLM-4.5V costs 22× less per token, which makes it the better buy when GPT-6 Astra's lead doesn't matter for your workload.
Which is cheaper, GLM-4.5V or GPT-6 Astra?
GLM-4.5V is cheaper. It lists at $0.60 per million input tokens and $1.80 per million output tokens; GPT-6 Astra lists at $10 and $50.
Is GLM-4.5V or GPT-6 Astra better for coding?
GPT-6 Astra scores higher on coding benchmarks: 73.7 versus 39.5 in the Noometry coding category.
Which has the bigger context window?
GPT-6 Astra does, with 1.05M tokens against 64K.
How many benchmarks do GLM-4.5V and GPT-6 Astra share?
14 benchmarks have published results for both models. GLM-4.5V has 15 scored results on Noometry and GPT-6 Astra has 56.