GLM 5.3 Flash vs GPT-6 Astra
Full profiles: GLM 5.3 Flash (Zhipu AI) and GPT-6 Astra (OpenAI): every benchmark, price and rank.
Updated · Data from 10 sources
As of 27 Oct 2026, GPT-6 Astra (max) leads on quality (70.3 vs 62.2 on the BenchLeader Index), but GLM 5.3 Flash costs 84× less, so it is the better buy unless you need the extra quality. GPT-6 Astra leads in every category we measure both in.
GLM 5.3 Flash benchmarks 
GPT-6 Astra benchmarks Which to pick for…
The pick's numbers come first. Category scores on the BenchLeader scale (50 is the average model, 15 points one standard deviation). High confidence means a lead of half a standard deviation or more; under one point is called either.
| Use case | Pick | Why | Confidence |
|---|---|---|---|
| Coding | GPT-6 Astra | 74 vs 63 category score; Arena WebDev 1774 vs 1608 | High |
| Agents & tool use | GPT-6 Astra | 64 vs 60 category score; Arena Agent 14 vs -1.8 | Medium |
| Maths | GPT-6 Astra | 72 vs 69 category score; Arena Maths 1487 vs 1504 | Medium |
| Reasoning | GPT-6 Astra | 76 vs 66 category score; CritPt 31.7% vs 15.4% | High |
| Knowledge | GPT-6 Astra | 79 vs 65 category score; AA-Omniscience 43.4 vs 7.5 | High |
| Long documents | Either | 65 vs 65 category score; AA-LCR 80.0% vs 80.7% | Either |
| Images and other media | GPT-6 Astra | 66 vs 64 category score; Arena Vision 1280 vs 1294 | Medium |
| Chat quality (human preference) | Either | 67 vs 67 on Arena votes | Either |
| Lowest cost | GLM 5.3 Flash | $0.238 vs $20.00 per 1M blended; agentic coding session $0.011 vs $0.800 | – |
| Fastest first answer | GLM 5.3 Flash | 30.2 s (GLM 5.3 Flash) vs 353.2 s (GPT-6 Astra (max)) to the first answer token | – |
| High volume on a budget | Neither flagship | A smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs | – |
Key differences
- GPT-6 Astra (max) leads on quality: 70.3 vs 62.2.
- GPT-6 Astra (max) is stronger in agents & tools, coding, composite, knowledge, long context, maths, multimodal, reasoning.
- GLM 5.3 Flash is 84× cheaper ($0.238 vs $20.00 per 1M blended).
At a glance
| Spec | GLM 5.3 Flash | GPT-6 Astra |
|---|---|---|
| Developer | Zhipu AI | OpenAI |
| Released | 20 Aug 2026 | 2 Sept 2026 |
| Context window (tokens) | 1,000,000 | 1,050,000 |
| Compared at | GLM 5.3 Flash | GPT-6 Astra (max) |
| Open weights | Yes | No |
| Price in / out / cached input per 1M | $0.150 / $0.500 / $0.034 | $10.00 / $50.00 / $1.00 |
Category by category
Category scores on the BenchLeader scale: 50 is the average evaluated model, 15 points is one standard deviation. A gap under one point is called level.
| Category | GLM 5.3 Flash | GPT-6 Astra (max) | Leader |
|---|---|---|---|
| Reasoning | 66.1 | 76.1 | GPT-6 Astra |
| Coding | 62.8 | 73.9 | GPT-6 Astra |
| Agents & tools | 59.9 | 64.2 | GPT-6 Astra |
| Maths | 68.7 | 71.8 | GPT-6 Astra |
| Knowledge | 64.5 | 79.3 | GPT-6 Astra |
| Human preference | 66.7 | 66.7 | level |
| Multimodal | 63.8 | 65.9 | GPT-6 Astra |
| Long context | 64.9 | 65.2 | level |
| Composite | 60.5 | 87.8 | GPT-6 Astra |
Where sources disagree
Across the 44 benchmarks both have, GLM 5.3 Flash leads 8, GPT-6 Astra leads 35 and 1 is a tie. Every overall index that rates both puts GPT-6 Astra ahead.
| Overall index | GLM 5.3 Flash | GPT-6 Astra (max) | Leader |
|---|---|---|---|
| BenchLeader Index | 62.2 | 70.3 | GPT-6 Astra |
| AA Intelligence Index v4.3.2 | 41.8 | 52.7 | GPT-6 Astra |
| Vals Index | 50.9† | 65.0 | GPT-6 Astra |
| Arena Text | 1476 | 1477 | GPT-6 Astra |
Each publisher runs its own tasks, harnesses and effort settings, so their overall verdicts can differ; BenchLeader reconciles them on one scale. † measured at another effort setting of the same model (the publisher did not test the one compared here).
Price, speed and latency
List prices blended at three input tokens per output token. Per-task prices use the token counts on the methodology page; suite cost is what Artificial Analysis paid to run its whole Intelligence Index.
| Measure | GLM 5.3 Flash | GPT-6 Astra (max) |
|---|---|---|
| Blended price per 1M tokens | $0.238 | $20.00 |
| Input / output per 1M | $0.150 / $0.500 | $10.00 / $50.00 |
| Cached input per 1M | $0.034 | $1.00 |
| Price per task: chat reply | $0.0002 | $0.019 |
| Price per task: summarize a 30-page report | $0.0021 | $0.150 |
| Price per task: code edit | $0.0017 | $0.135 |
| Price per task: agentic coding session | $0.011 | $0.800 |
| Price per task: structured extraction | $0.0004 | $0.030 |
| Cost to run a full benchmark suite | $0.284 | $3.26 |
| Output speed | 73 tok/s | 68 tok/s |
| Time to first answer | 30.2 s | 353.2 s |
| Time to a complete answer | 37.0 s | 360.5 s |
Per token is not per task: on Artificial Analysis' full suite GLM 5.3 Flash cost $0.284 and GPT-6 Astra (max) $3.26, because GLM 5.3 Flash used about 7.3× as many tokens for the same work, which a per-token price hides.
Each model is compared at its best-ranked reasoning-effort setting. GLM 5.3 Flash scores 62.2 and takes 30 s to its first answer token; GLM 5.3 Flash (max) scores 54.2 in 30 s. GPT-6 Astra (max) scores 70.3 and takes 353 s to its first answer token; GPT-6 Astra (high) scores 69.5 in 62 s, GPT-6 Astra (xhigh) scores 68.8 in 175 s. Higher effort means more thinking before the answer: better scores, longer waits and more tokens billed. How reasoning effort changes scores and cost.
Every benchmark
142 benchmarks, grouped by category; each name links to the board and its leaders. † measured at another effort setting of the same model, shown for reference and not counted as a win.
| Metric | Publisher | GLM 5.3 Flash | GPT-6 Astra (max) |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 62.2 | 70.3 |
| Agents & tools score | BenchLeader | 59.9 | 64.2 |
| Coding score | BenchLeader | 62.8 | 73.9 |
| Composite score | BenchLeader | 60.5 | 87.8 |
| Human preference score | BenchLeader | 66.7 | 66.7 |
| Knowledge score | BenchLeader | 64.5 | 79.3 |
| Long context score | BenchLeader | 64.9 | 65.2 |
| Maths score | BenchLeader | 68.7 | 71.8 |
| Multimodal score | BenchLeader | 63.8 | 65.9 |
| Reasoning score | BenchLeader | 66.1 | 76.1 |
| Blended price $/M | BenchLeader | $0.238 | $20.00 |
| Output speed | BenchLeader | 73 tok/s | 68 tok/s |
| Time to first answer | BenchLeader | 30.2 s | 353.2 s |
| Context window | BenchLeader | 1,000,000 tokens | 1,050,000 tokens |
| Reasoning | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | 90.2%† | 95.8% |
| Humanity's Last Exam | Scale AI / CAIS | – | 54.8%† |
| Arena Hard Prompts | Arena (LMArena) | 1501 | 1501 |
| LiveBench 2.0 Reasoning | LiveBench | 77.6% | 92.7%† |
| LiveBench 1.0 Reasoning | LiveBench | 77.6% | 92.7% |
| GPQA Diamond (AA) | Artificial Analysis | 91.2% | 96.1% |
| Humanity's Last Exam (AA) | Artificial Analysis | 39.9% | 54.7% |
| GPQA Diamond (Vals) | Vals AI | 86.4%† | – |
| ARC-AGI-1 | ARC Prize | 91.0%† | 97.5% |
| ARC-AGI-2 | ARC Prize | 65.8%† | 95.0% |
| ARC-AGI-3 | ARC Prize | – | 98.6% |
| CritPt | Artificial Analysis | 15.4% | 31.7% |
| MysteryMechanism | Vals AI | – | 53.1% |
| Chess Puzzles | Epoch AI Benchmarking Hub | 21.0%† | 74.0% |
| Mystery Game Puzzles | Epoch AI Benchmarking Hub | 8.0%† | 84.0% |
| LMCA | Epoch AI Benchmarking Hub | – | 64.1%† |
| DTBench | Epoch AI Benchmarking Hub | – | 97.1%† |
| BTF-3 | Epoch AI Benchmarking Hub | 14.9% | 13.5%† |
| Coding | |||
| SciCode | SciCode | 51.6% | 56.5% |
| WeirdML | WeirdML | – | 93.3% |
| FrontierCode | Cognition | 31.8%† | 53.3% |
| GSO-Bench | GSO-Bench | – | 79.4%† |
| Arena Coding | Arena (LMArena) | 1525 | 1543 |
| Arena WebDev | Arena (LMArena) | 1608 | 1774 |
| LiveBench 2.0 Coding | LiveBench | 80.6% | 93.7%† |
Show all 142 benchmarks
| Metric | Publisher | GLM 5.3 Flash | GPT-6 Astra (max) |
|---|---|---|---|
| Coding | |||
| LiveBench 1.0 Coding | LiveBench | 79.0% | 80.4% |
| LiveCodeBench | Vals AI | 80.5%† | – |
| IOI | Vals AI | 52.5%† | 100.0% |
| SWE-bench (Vals) | Vals AI | 92.0%† | – |
| Code Migration | Vals AI | 20.5%† | 67.7% |
| Code Migration v1.1 | Vals AI | 51.6%† | 80.6% |
| ProgramBench | Vals AI | 0.0%† | 5.5% |
| SWE-bench Multimodal | Vals AI | – | 43.3% |
| Vibe Code Bench 1-100 | Vals AI | 16.0%† | 27.6% |
| Vibe Code Bench v1.1 | Vals AI | 30.8%† | 89.6% |
| SWE Atlas: Codebase QnA | Scale AI SEAL | – | 59.1%† |
| SWE Atlas: Refactoring | Scale AI SEAL | – | 59.0%† |
| SWE Atlas: Test Writing | Scale AI SEAL | – | 50.7%† |
| MirrorCode | Epoch AI Benchmarking Hub | – | 46.7%† |
| Surface Evolver Bench | Epoch AI Benchmarking Hub | 52.5%† | – |
| DeepSWE v1.1 | Epoch AI Benchmarking Hub | 63.4%† | 73.2% |
| CursorBench | Cursor | 36.8%† | – |
| ALE-Bench | Epoch AI Benchmarking Hub | 303.6† | 2951.3 |
| FrontierSWE | Epoch AI Benchmarking Hub | 18.1%† | 65.5% |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 32.8% | 59.1% |
| Terminal-Bench 2.1 (AA) | Artificial Analysis | 84.3% | 88.4% |
| Startrise Frontend | Startrise | – | 79.9%† |
| FrontierCode 1.1 (AA) | Artificial Analysis | 41.0% | 47.0% |
| Agents & tools | |||
| Terminal-Bench | Terminal-Bench | 35.8%† | 58.2% |
| Remote Labor Index | Scale AI / CAIS | – | 20.8%† |
| APEX-Agents | Mercor | 52.8% | 64.7%† |
| Arena Agent | Arena (LMArena) | -1.8 | 14 |
| LiveBench Long Agentic | LiveBench | 19.0% | 77.5%† |
| LiveBench Browser | LiveBench | 59.8% | 91.3%† |
| LiveBench Agentic Coding | LiveBench | 56.8% | 57.3% |
| Terminal-Bench 2.1 (Vals) | Vals AI | 62.9%† | 87.3% |
| HiL-Bench | Scale AI SEAL | – | 65.3%† |
| GDPval-AA v2.1 | Artificial Analysis | 57.1% | 53.7% |
| τ³-Banking (AA) | Artificial Analysis | 47.2% | 41.4% |
| ITBench SRE (AA) | Artificial Analysis | 51.2% | 48.6% |
| Analyst Agent (AA) | Artificial Analysis | 25.0% | 51.3% |
| CUA-bench | Vals AI | – | 19.2% |
| CyberBench | Vals AI | 67.9%† | 41.1% |
| Finance Agent v2 | Vals AI | 57.9%† | 53.5% |
| Finance Agent v2.1 | Vals AI | 60.6%† | 60.4% |
| Harvey's Legal Agent Benchmark v1.1 | Vals AI | 10.4%† | 7.1% |
| Harvey's Legal Agent Benchmark v1 | Vals AI | 6.7%† | 5.4% |
| Legal Research Bench | Vals AI | 45.2%† | 39.4% |
| Robo-bench | Vals AI | – | 23.3% |
| SkillsBench | Vals AI | 40.2%† | – |
| SREBench | Vals AI | 0.8%† | 56.9% |
| Tax Agent Bench | Vals AI | 31.6%† | 20.7% |
| Terminal-Bench 4.0 (Vals) | Vals AI | 37.4%† | 59.6% |
| Terminal-Bench Science | Vals AI | 2.9%† | 62.9% |
| Time Horizon Index: KSP | Vals AI | – | 90.5% |
| Arena Search | Arena (LMArena) | – | 1246 |
| BALROG | BALROG | – | 68.3% |
| Vending-Bench 2 | Andon Labs | – | 15514.7† |
| AutomationBench | Artificial Analysis | 60.4% | 68.5% |
| GDP.pdf (AA) | Artificial Analysis | 15.4% | 31.0% |
| Harvey LAB-AA v1.1 | Harvey | 0.8% | 8.6% |
| EnterpriseOps-Gym | Artificial Analysis | 33.2% | – |
| AA-Briefcase v1.1 | Artificial Analysis | 1450 | 1568 |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | 4.8% | 63.3% |
| CWE-Bench (AA) | Artificial Analysis | 55.8% | 63.3% |
| DeepSec-Bench (AA) | Artificial Analysis | 19.9% | 36.9% |
| CyberGym end-to-end (AA) | Artificial Analysis | 74.0% | 0.0% |
| FACTS Search | – | 89.9%† | |
| Maths | |||
| FrontierMath Tiers 1–3 | Epoch AI Benchmarking Hub | 55.8%† | 93.7% |
| FrontierMath Tier 4 | Epoch AI Benchmarking Hub | 17.1%† | 97.6% |
| OTIS Mock AIME | Epoch AI Benchmarking Hub | 93.9%† | 100.0% |
| ProofBench | Vals AI | 21.0%† | 99.0%† |
| LiveBench Mathematics | LiveBench | 81.2% | 96.8% |
| Arena Maths | Arena (LMArena) | 1504 | 1487 |
| Knowledge | |||
| SimpleQA Verified | Epoch AI Benchmarking Hub | – | 75.6% |
| LiveBench 2.0 Data Analysis | LiveBench | 76.4% | 83.0%† |
| LiveBench Research | LiveBench | 70.1% | 85.2%† |
| LiveBench 1.0 Data Analysis | LiveBench | 76.4% | 83.0% |
| AA-Omniscience | Artificial Analysis | 7.5 | 43.4 |
| MMLU-Pro | Vals AI | 86.1%† | – |
| LegalBench | Vals AI | 83.9%† | – |
| TaxEval | Vals AI | 75.6%† | – |
| PRBench Finance | Scale AI SEAL | – | 47.5%† |
| PRBench Legal | Scale AI SEAL | – | 48.4%† |
| BioMysteryBench | Vals AI | – | 79.3% |
| Excel Modeling Benchmark | Vals AI | 55.9%† | 71.7% |
| Excel Modeling Benchmark v1.1 | Vals AI | 66.3%† | 72.7% |
| MedCode | Vals AI | – | 48.5% |
| MedScribe | Vals AI | 88.9%† | 87.9% |
| DrugDiscoveryBench | Scale AI SEAL | – | 68.7% |
| EBR-bench | Epoch AI Benchmarking Hub | – | 76.2% |
| GDP.pdf | Epoch AI Benchmarking Hub | 14.0%† | 34.2% |
| AA-Omniscience: accuracy | Artificial Analysis | 27.5% | 62.6% |
| AA-Omniscience: non-hallucination | Artificial Analysis | 72.4% | 48.7% |
| Instruction following | |||
| LiveBench Language | LiveBench | 77.3% | 89.4% |
| LiveBench Instruction Following | LiveBench | 52.8% | 75.6% |
| Arena Instruction Following | Arena (LMArena) | 1473 | 1473 |
| Human preference | |||
| Arena Text | Arena (LMArena) | 1476 | 1477 |
| Arena Creative Writing | Arena (LMArena) | 1434 | 1450 |
| Arena Multi-turn | Arena (LMArena) | 1476 | 1485 |
| Arena Longer Queries | Arena (LMArena) | 1481 | 1489 |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | 1294 | 1280 |
| LiveBench Design | LiveBench | 77.2% | 82.0%† |
| MMMU-Pro | Artificial Analysis | – | 86.9% |
| MMMU-Pro (Vals) | Vals AI | 86.0%† | – |
| SAGE | Vals AI | – | 46.4% |
| Blueprint-Bench 2 | Epoch AI Benchmarking Hub | – | 49.7%† |
| AA-3D | Artificial Analysis | 937 | 1164 |
| Long context | |||
| AA-LCR | Artificial Analysis | 80.0% | 80.7% |
| Arena Document | Arena (LMArena) | – | 1473 |
| MLCR | Artificial Analysis | 11.7% | 41.7% |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | 153.1 | 168.8† |
| LiveBench 2.0 | LiveBench | 62.1% | 85.6%† |
| LiveBench 1.0 | LiveBench | 71.6% | 82.2% |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 41.8 | 52.7 |
| Vals Index | Vals AI | 50.9† | 65.0 |
| Vals Finance Index | Vals AI | 55.8%† | 55.6% |
| Vals RSI Index | Vals AI | – | 27.1% |
| Vals Science Index | Vals AI | – | 65.1% |
| Vals Router Index | Vals AI | 50.9% | 65.0%† |
| AA Cyber Index | Artificial Analysis | 49.9 | 33.4 |
| Metadata | |||
| AA Openness Index | Artificial Analysis | 44.4 | – |
Data as of 2026-10-27. Best configuration of each model; every score links to its source on the model pages.
What changed
- GLM 5.3 Flash: GLM 5.3 Flash cached input price moved from $0.030/M to $0.034/M (15% up).
- GLM 5.3 Flash: BenchLeader Index 62.9 → 62.3.
- GLM 5.3 Flash: 3 score revisions; GLM 5.3 Flash cached input price moved from $0.034/M to $0.030/M (13% down).
- GLM 5.3 Flash: GLM 5.3 Flash cached input price moved from $0.030/M to $0.034/M (15% up).
- GPT-6 Astra (max): 1 score revision.
- GLM 5.3 Flash: GLM 5.3 Flash cached input price moved from $0.034/M to $0.030/M (13% down).
- GLM 5.3 Flash: 1 new result (Vals Router Index).
- GPT-6 Astra (max): 9 new results (Code Migration v1.1, Excel Modeling Benchmark v1.1, Finance Agent v2.1…).
- GLM 5.3 Flash: 1 score revision.
- GPT-6 Astra (max): 1 score revision.
Related comparisons
- GLM 5.3 Flash vs Claude Opus 5.5
- GLM 5.3 Flash vs Claude Fable 5.1
- GLM 5.3 Flash vs Muse Wonder Full
- GPT-6 Astra vs Claude Opus 5.5
- GPT-6 Astra vs Claude Fable 5.1
- GPT-6 Astra vs Muse Wonder Full
- GPT-6 Astra vs GPT-6.1 Sol
- GLM 5.3 vs GPT-6 Astra
Rankings: Best LLM overall · Best LLM for coding · Best LLM for agents · Best cheap LLMs
GLM 5.3 Flash vs GPT-6 Astra: questions
- Is GLM 5.3 Flash better than GPT-6 Astra?
- GPT-6 Astra (max) leads on quality: 70.3 vs 62.2. The BenchLeader Index combines every independent quality benchmark; GPT-6 Astra (max) is ahead overall as of 2026-10-27, but check the category scores for your use.
- Which is cheaper, GLM 5.3 Flash or GPT-6 Astra, and will that change?
- GLM 5.3 Flash is cheaper today: $0.238 against $20.00 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
- Is GLM 5.3 Flash better than GPT-6 Astra for coding?
- GPT-6 Astra scores higher in coding (74 vs 63 on the category index, where 50 is average). On Arena WebDev, the top shared coding benchmark, it is 1608 for GLM 5.3 Flash against 1774 for GPT-6 Astra.
- Is GLM 5.3 Flash better than GPT-6 Astra for agents and tool use?
- GPT-6 Astra scores higher in agentic tasks (64 vs 60 on the category index, where 50 is average). On τ³-Banking (AA), the top shared agents & tools benchmark, it is 47.2% for GLM 5.3 Flash against 41.4% for GPT-6 Astra.
- Which is better for long documents?
- They are level on long-context benchmarks (65 vs 65). Both take about 1.1M tokens: GLM 5.3 Flash 1,000,000, GPT-6 Astra 1,050,000.
- Which has lower latency?
- GLM 5.3 Flash answers first: 30.2 s against 353.2 s to the first answer token, as Artificial Analysis measures it. Reasoning effort drives most of this: a higher setting thinks longer before it answers.
- Which is faster, GLM 5.3 Flash or GPT-6 Astra?
- GLM 5.3 Flash streams faster: 73 against 68 output tokens per second.
- Do independent benchmarks agree on GLM 5.3 Flash vs GPT-6 Astra?
- Across the 44 benchmarks both have, GLM 5.3 Flash leads 8, GPT-6 Astra leads 35 and 1 is a tie. Every overall index that rates both puts GPT-6 Astra ahead.
- When was GPT-6 Astra released?
- GPT-6 Astra was released on 2 Sept 2026.