Gemini 3.5 Flash vs GPT-6 Astra
Full profiles: Gemini 3.5 Flash (Google) and GPT-6 Astra (OpenAI): every benchmark, price and rank.
Updated · Data from 10 sources
As of 24 Oct 2026, GPT-6 Astra (max) leads on quality (70.4 vs 63.1 on the BenchLeader Index), but Gemini 3.5 Flash (high) costs 5.9× less, so it is the better buy unless you need the extra quality. GPT-6 Astra leads in every category we measure both in.
Gemini 3.5 Flash benchmarks high Google
GPT-6 Astra benchmarks max OpenAIOpen in the comparison toolWhich 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 60 category score; Arena WebDev 1775 vs 1499 | High |
| Agents & tool use | Either | 63 vs 64 category score; τ³-Banking (AA) 32.2% vs 41.4% | Either |
| Maths | GPT-6 Astra | 72 vs 58 category score; FrontierMath Tiers 1–3 93.7% vs 62.8% | High |
| Reasoning | GPT-6 Astra | 76 vs 61 category score; GPQA Diamond 95.8% vs 92.8% | High |
| Knowledge | GPT-6 Astra | 80 vs 65 category score; SimpleQA Verified 75.6% vs 66.2% | High |
| Long documents | GPT-6 Astra | 65 vs 62 category score; AA-LCR 80.7% vs 73.3% | Medium |
| Images and other media | Either | 67 vs 66 category score; Arena Vision 1308 vs 1280 | Either |
| Chat quality (human preference) | Either | 67 vs 67 on Arena votes | Either |
| Lowest cost | Gemini 3.5 Flash | $3.38 vs $20.00 per 1M blended; agentic coding session $0.126 vs $0.800 | – |
| Fastest first answer | Gemini 3.5 Flash | 14.3 s (Gemini 3.5 Flash (high)) vs 247.0 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: 70.4 vs 63.1.
- Gemini 3.5 Flash (high) is stronger in human preference, instruction following, multimodal.
- GPT-6 Astra (max) is stronger in agents & tools, coding, composite, knowledge, long context, maths, reasoning.
- Gemini 3.5 Flash (high) is 5.9× cheaper ($3.38 vs $20.00 per 1M blended).
- Gemini 3.5 Flash (high) streams 3.8× faster (222 vs 58 tokens per second).
At a glance
| Spec | Gemini 3.5 Flash | GPT-6 Astra |
|---|---|---|
| Developer | OpenAI | |
| Released | 19 May 2026 | 2 Sept 2026 |
| Context window (tokens) | 1,048,576 | 1,050,000 |
| Compared at | Gemini 3.5 Flash (high) | GPT-6 Astra (max) |
| Open weights | No | No |
| Price in / out / cached input per 1M | $1.50 / $9.00 / $0.150 | $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 | Gemini 3.5 Flash (high) | GPT-6 Astra (max) | Leader |
|---|---|---|---|
| Reasoning | 61.3 | 76.2 | GPT-6 Astra |
| Coding | 59.9 | 74.0 | GPT-6 Astra |
| Agents & tools | 63.4 | 64.2 | level |
| Maths | 58.3 | 71.8 | GPT-6 Astra |
| Knowledge | 65.2 | 79.5 | GPT-6 Astra |
| Instruction following | 74.3 | – | not comparable |
| Human preference | 66.8 | 66.7 | level |
| Multimodal | 66.6 | 66.0 | level |
| Long context | 61.6 | 65.4 | GPT-6 Astra |
| Composite | 65.9 | 88.4 | GPT-6 Astra |
Where sources disagree
Across the 64 benchmarks both have, Gemini 3.5 Flash leads 8, GPT-6 Astra leads 56. Independent sources split: Gemini 3.5 Flash leads on Arena Text; GPT-6 Astra leads on BenchLeader Index and AA Intelligence Index v4.3.2.
| Overall index | Gemini 3.5 Flash (high) | GPT-6 Astra (max) | Leader |
|---|---|---|---|
| BenchLeader Index | 63.1 | 70.4 | GPT-6 Astra |
| AA Intelligence Index v4.3.2 | 32.6 | 52.7 | GPT-6 Astra |
| Arena Text | 1477 | 1477 | Gemini 3.5 Flash |
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 | Gemini 3.5 Flash (high) | GPT-6 Astra (max) |
|---|---|---|
| Blended price per 1M tokens | $3.38 | $20.00 |
| Input / output per 1M | $1.50 / $9.00 | $10.00 / $50.00 |
| Cached input per 1M | $0.150 | $1.00 |
| Price per task: chat reply | $0.0033 | $0.019 |
| Price per task: summarize a 30-page report | $0.023 | $0.150 |
| Price per task: code edit | $0.022 | $0.135 |
| Price per task: agentic coding session | $0.126 | $0.800 |
| Price per task: structured extraction | $0.0048 | $0.030 |
| Cost to run a full benchmark suite | $3.69 | $3.26 |
| Output speed | 222 tok/s | 58 tok/s |
| Time to first answer | 14.3 s | 247.0 s |
| Time to a complete answer | 16.6 s | 255.6 s |
Per token is not per task: on Artificial Analysis' full suite Gemini 3.5 Flash (high) cost $3.69 and GPT-6 Astra (max) $3.26, because Gemini 3.5 Flash (high) used about 6.7× as many tokens for the same work, which a per-token price hides.
Each model is compared at its best-ranked reasoning-effort setting. Gemini 3.5 Flash (high) scores 63.1 and takes 14 s to its first answer token; Gemini 3.5 Flash (medium) scores 62.8 in 13 s, Gemini 3.5 Flash (minimal) scores 56.4 in 1 s. GPT-6 Astra (max) scores 70.4 and takes 247 s to its first answer token; GPT-6 Astra (high) scores 69.6 in 35 s, GPT-6 Astra (xhigh) scores 68.9 in 130 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
160 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 | Gemini 3.5 Flash (high) | GPT-6 Astra (max) |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 63.1 | 70.4 |
| Agents & tools score | BenchLeader | 63.4 | 64.2 |
| Coding score | BenchLeader | 59.9 | 74.0 |
| Composite score | BenchLeader | 65.9 | 88.4 |
| Human preference score | BenchLeader | 66.8 | 66.7 |
| Instruction following score | BenchLeader | 74.3 | – |
| Knowledge score | BenchLeader | 65.2 | 79.5 |
| Long context score | BenchLeader | 61.6 | 65.4 |
| Maths score | BenchLeader | 58.3 | 71.8 |
| Multimodal score | BenchLeader | 66.6 | 66.0 |
| Reasoning score | BenchLeader | 61.3 | 76.2 |
| Blended price $/M | BenchLeader | $3.38 | $20.00 |
| Output speed | BenchLeader | 222 tok/s | 58 tok/s |
| Time to first answer | BenchLeader | 14.3 s | 247.0 s |
| Context window | BenchLeader | 1,048,576 tokens | 1,050,000 tokens |
| Reasoning | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | 92.8% | 95.8% |
| Humanity's Last Exam | Scale AI / CAIS | – | 54.8%† |
| SimpleBench | SimpleBench | 76.7%† | – |
| Arena Hard Prompts | Arena (LMArena) | 1495 | 1501 |
| LiveBench 2.0 Reasoning | LiveBench | – | 92.7%† |
| LiveBench 1.0 Reasoning | LiveBench | 82.0% | 92.7% |
| GPQA Diamond (AA) | Artificial Analysis | 92.2% | 96.1% |
| Humanity's Last Exam (AA) | Artificial Analysis | 42.7% | 54.7% |
| GPQA Diamond (Vals) | Vals AI | 92.7% | – |
| EnigmaEval | Scale AI SEAL | 25.4% | – |
| ARC-AGI-1 | ARC Prize | 92.5% | 97.5% |
| ARC-AGI-2 | ARC Prize | 72.1% | 95.0% |
| ARC-AGI-3 | ARC Prize | – | 98.6% |
| CritPt | Artificial Analysis | 13.1% | 31.7% |
| MysteryMechanism | Vals AI | – | 53.1% |
| Chess Puzzles | Epoch AI Benchmarking Hub | 45.0% | 74.0% |
| Mystery Game Puzzles | Epoch AI Benchmarking Hub | 32.0% | 84.0% |
| LMCA | Epoch AI Benchmarking Hub | 47.1% | 64.1%† |
| DTBench | Epoch AI Benchmarking Hub | 94.7% | 97.1%† |
| ForecastBench | ForecastBench | 64.3%† | – |
| BTF-3 | Epoch AI Benchmarking Hub | – | 13.5%† |
| Coding | |||
| SWE-bench Verified (Epoch) | Epoch AI Benchmarking Hub | 79.3% | – |
| SciCode | SciCode | 53.9% | 56.5% |
| WeirdML | WeirdML | 62.6% | 93.3% |
| FrontierCode | Cognition | – | 53.3% |
Show all 160 benchmarks
| Metric | Publisher | Gemini 3.5 Flash (high) | GPT-6 Astra (max) |
|---|---|---|---|
| Coding | |||
| GSO-Bench | GSO-Bench | – | 79.4%† |
| Arena Coding | Arena (LMArena) | 1508 | 1543 |
| Arena WebDev | Arena (LMArena) | 1499 | 1775 |
| LiveBench 2.0 Coding | LiveBench | – | 93.7%† |
| LiveBench 1.0 Coding | LiveBench | 78.2% | 80.4% |
| LiveCodeBench | Vals AI | 87.6% | – |
| IOI | Vals AI | – | 100.0% |
| SWE-bench (Vals) | Vals AI | 78.8% | – |
| Code Migration | Vals AI | 26.8% | 67.7% |
| Code Migration v1.1 | Vals AI | – | 80.6% |
| ProgramBench | Vals AI | 0.0% | 5.5% |
| SWE-bench Multimodal | Vals AI | – | 43.3% |
| Vibe Code Bench 1-100 | Vals AI | – | 27.6% |
| Vibe Code Bench v1.1 | Vals AI | 48.7% | 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 | 58.1%† | – |
| DeepSWE v1.1 | Epoch AI Benchmarking Hub | 36.1% | 73.2% |
| ALE-Bench | Epoch AI Benchmarking Hub | 911.0 | 2951.3 |
| FrontierSWE | Epoch AI Benchmarking Hub | – | 65.5% |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 6.6% | 59.1% |
| Terminal-Bench 2.1 (AA) | Artificial Analysis | 78.7% | 88.4% |
| Startrise Frontend | Startrise | – | 79.9%† |
| FrontierCode 1.1 (AA) | Artificial Analysis | – | 47.0% |
| Agents & tools | |||
| Terminal-Bench | Terminal-Bench | – | 58.2% |
| Remote Labor Index | Scale AI / CAIS | – | 20.8%† |
| APEX-Agents | Mercor | 27.5%† | 64.7%† |
| Arena Agent | Arena (LMArena) | – | 14.7 |
| LiveBench Long Agentic | LiveBench | – | 82.4%† |
| LiveBench Browser | LiveBench | – | 84.0%† |
| LiveBench Agentic Coding | LiveBench | 49.0% | 57.3% |
| Terminal-Bench Hard | Artificial Analysis | 40.9% | – |
| τ²-Bench Telecom (AA) | Artificial Analysis | 95.3% | – |
| Terminal-Bench 2.1 (Vals) | Vals AI | 74.2% | 87.3% |
| MCP Atlas | Scale AI SEAL | 83.6% | – |
| HiL-Bench | Scale AI SEAL | 27.7%† | 65.3%† |
| GDPval-AA v2.1 | Artificial Analysis | – | 53.7% |
| τ³-Banking (AA) | Artificial Analysis | 32.2% | 41.4% |
| ITBench SRE (AA) | Artificial Analysis | 40.4% | 48.6% |
| Analyst Agent (AA) | Artificial Analysis | 45.0% | 51.3% |
| APEX-Agents (AA) | Artificial Analysis | 47.0% | – |
| CUA-bench | Vals AI | – | 19.2% |
| CyberBench | Vals AI | – | 41.1% |
| Finance Agent v2 | Vals AI | 57.9% | 53.5% |
| Finance Agent v2.1 | Vals AI | – | 60.4% |
| Harvey's Legal Agent Benchmark v1.1 | Vals AI | – | 7.1% |
| Harvey's Legal Agent Benchmark v1 | Vals AI | 2.5% | 5.4% |
| Legal Research Bench | Vals AI | 30.8% | 39.4% |
| Robo-bench | Vals AI | – | 23.3% |
| SkillsBench | Vals AI | 52.7% | – |
| SREBench | Vals AI | – | 56.9% |
| Tax Agent Bench | Vals AI | 21.6% | 20.7% |
| Terminal-Bench 2.0 (Vals) | Vals AI | 67.4% | – |
| Terminal-Bench 4.0 (Vals) | Vals AI | 6.1% | 59.6% |
| Terminal-Bench Science | Vals AI | 5.7% | 62.9% |
| Time Horizon Index: KSP | Vals AI | – | 90.5% |
| Web Search Index | Vals AI | 37.0% | – |
| Arena Search | Arena (LMArena) | – | 1246 |
| BALROG | BALROG | – | 68.3% |
| GBAEval | Epoch AI Benchmarking Hub | 6.7%† | – |
| Vending-Bench 2 | Andon Labs | 5396.4† | 15514.7† |
| AutomationBench | Artificial Analysis | 42.1% | 68.5% |
| GDP.pdf (AA) | Artificial Analysis | 19.8% | 31.0% |
| Harvey LAB-AA v1.1 | Harvey | – | 8.6% |
| EnterpriseOps-Gym | Artificial Analysis | 50.1% | – |
| AA-Briefcase v1.1 | Artificial Analysis | 874 | 1568 |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | – | 63.3% |
| CWE-Bench (AA) | Artificial Analysis | – | 63.3% |
| DeepSec-Bench (AA) | Artificial Analysis | – | 36.9% |
| CyberGym end-to-end (AA) | Artificial Analysis | – | 0.0% |
| FACTS Search | 64.3%† | 89.9%† | |
| Maths | |||
| FrontierMath Tiers 1–3 | Epoch AI Benchmarking Hub | 62.8% | 93.7% |
| FrontierMath Tier 4 | Epoch AI Benchmarking Hub | 26.8% | 97.6% |
| OTIS Mock AIME | Epoch AI Benchmarking Hub | 95.6% | 100.0% |
| ProofBench | Vals AI | 31.0% | 99.0%† |
| LiveBench Mathematics | LiveBench | 88.2% | 96.8% |
| AIME 2026 | MathArena | 95.0%† | – |
| HMMT February 2026 | MathArena | 95.5%† | – |
| MathArena Apex | MathArena | 32.3%† | – |
| Arena Maths | Arena (LMArena) | 1497 | 1487 |
| Knowledge | |||
| SimpleQA Verified | Epoch AI Benchmarking Hub | 66.2% | 75.6% |
| LiveBench 2.0 Data Analysis | LiveBench | – | 83.0%† |
| LiveBench Research | LiveBench | – | 85.2%† |
| LiveBench 1.0 Data Analysis | LiveBench | 64.9% | 83.0% |
| AA-Omniscience | Artificial Analysis | 21.2 | 43.4 |
| MMLU-Pro | Vals AI | 89.5% | – |
| LegalBench | Vals AI | 83.6% | – |
| CorpFin | Vals AI | 64.7% | – |
| TaxEval | Vals AI | 74.4% | – |
| PRBench Finance | Scale AI SEAL | – | 47.5%† |
| PRBench Legal | Scale AI SEAL | – | 48.4%† |
| BioMysteryBench | Vals AI | – | 79.3% |
| Excel Modeling Benchmark | Vals AI | 63.5% | 71.7% |
| Excel Modeling Benchmark v1.1 | Vals AI | – | 72.7% |
| MedCode | Vals AI | 55.8% | 48.5% |
| MedScribe | Vals AI | 76.6% | 87.9% |
| Public Benefits Bench v1.1 | Vals AI | 59.5% | – |
| DrugDiscoveryBench | Scale AI SEAL | 50.0% | 68.7% |
| EBR-bench | Epoch AI Benchmarking Hub | 4.8% | 76.2% |
| GDP.pdf | Epoch AI Benchmarking Hub | 14.0%† | 34.2% |
| AA-Omniscience: accuracy | Artificial Analysis | 51.4% | 62.6% |
| AA-Omniscience: non-hallucination | Artificial Analysis | 37.8% | 48.7% |
| Instruction following | |||
| LiveBench Language | LiveBench | 84.6% | 89.4% |
| LiveBench Instruction Following | LiveBench | 75.6% | 75.6% |
| IFBench | Artificial Analysis | 76.3% | – |
| Arena Instruction Following | Arena (LMArena) | 1465 | 1473 |
| Human preference | |||
| Arena Text | Arena (LMArena) | 1477 | 1477 |
| EQ-Bench 4 | EQ-Bench | 1087† | – |
| Arena Creative Writing | Arena (LMArena) | 1464 | 1450 |
| Arena Multi-turn | Arena (LMArena) | 1480 | 1485 |
| Arena Longer Queries | Arena (LMArena) | 1482 | 1489 |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | 1308 | 1280 |
| LiveBench Design | LiveBench | – | 82.0%† |
| MMMU-Pro | Artificial Analysis | 84.3% | 86.9% |
| MMMU-Pro (Vals) | Vals AI | 88.3% | – |
| MortgageTax | Vals AI | 68.1% | – |
| SAGE | Vals AI | 49.9% | 46.4% |
| Blueprint-Bench 2 | Epoch AI Benchmarking Hub | 33.6%† | 49.7%† |
| AA-3D | Artificial Analysis | – | 1166 |
| Long context | |||
| AA-LCR | Artificial Analysis | 73.3% | 80.7% |
| Arena Document | Arena (LMArena) | 1453 | 1473 |
| MLCR | Artificial Analysis | 10.0% | 41.7% |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | 154.9† | 168.8† |
| LiveBench 2.0 | LiveBench | – | 86.1%† |
| LiveBench 1.0 | LiveBench | 74.6% | 82.2% |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 32.6 | 52.7 |
| Vals Index | Vals AI | – | 65.0 |
| Vals Finance Index | Vals AI | – | 55.6% |
| Vals RSI Index | Vals AI | – | 27.1% |
| Vals Science Index | Vals AI | – | 65.1% |
| Vals Multimodal Index | Vals AI | 62.9% | – |
| Vals Router Index | Vals AI | – | 65.0%† |
| AA Cyber Index | Artificial Analysis | – | 33.4 |
Data as of 2026-10-24. Best configuration of each model; every score links to its source on the model pages.
What changed
- Gemini 3.5 Flash (high): 1 new result (Harvey's Legal Agent Benchmark v1).
- GPT-6 Astra (max): 9 new results (Code Migration v1.1, Excel Modeling Benchmark v1.1, Finance Agent v2.1…).
- Gemini 3.5 Flash (high): 1 score revision.
- GPT-6 Astra (max): 1 score revision.
- GPT-6 Astra (max): 4 score revisions.
- Gemini 3.5 Flash (high): BenchLeader Index 62.5 → 63.2.
- GPT-6 Astra (max): BenchLeader Index 69.8 → 70.6.
- Gemini 3.5 Flash (high): 3 new results (GDPval-AA v2.1, AA-Omniscience: accuracy, AA-Omniscience: non-hallucination).
- Gemini 3.5 Flash (high): 1 score revision.
- GPT-6 Astra (max): 1 score revision.
Related comparisons
- Gemini 3.5 Flash vs Claude Opus 5.5
- Gemini 3.5 Flash vs Gemini 4 Argon
- Gemini 3.5 Flash vs Muse Wonder Full
- GPT-6 Astra vs Claude Opus 5.5
- GPT-6 Astra vs Gemini 4 Argon
- GPT-6 Astra vs Muse Wonder Full
- GPT-6 Astra vs Claude Fable 5.1
- Gemini 3.1 Pro vs GPT-6 Astra
Rankings: Best LLM overall · Best LLM for coding · Best LLM for agents · Best cheap LLMs
Gemini 3.5 Flash vs GPT-6 Astra: questions
- Is Gemini 3.5 Flash better than GPT-6 Astra?
- GPT-6 Astra (max) leads on quality: 70.4 vs 63.1. The BenchLeader Index combines every independent quality benchmark; GPT-6 Astra (max) is ahead overall as of 2026-10-24, but check the category scores for your use.
- Which is cheaper, Gemini 3.5 Flash or GPT-6 Astra, and will that change?
- Gemini 3.5 Flash is cheaper today: $3.38 against $20.00 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
- Is Gemini 3.5 Flash better than GPT-6 Astra for coding?
- GPT-6 Astra scores higher in coding (74 vs 60 on the category index, where 50 is average). On Arena WebDev, the top shared coding benchmark, it is 1499 for Gemini 3.5 Flash against 1775 for GPT-6 Astra.
- Is Gemini 3.5 Flash better than GPT-6 Astra for agents and tool use?
- They are level in agentic tasks (63 vs 64 on the category index). On τ³-Banking (AA), the top shared agents & tools benchmark, it is 32.2% for Gemini 3.5 Flash against 41.4% for GPT-6 Astra.
- Which is better for long documents?
- GPT-6 Astra scores higher on long-context benchmarks (65 vs 62). Both take about 1.1M tokens: Gemini 3.5 Flash 1,048,576, GPT-6 Astra 1,050,000.
- Which has lower latency?
- Gemini 3.5 Flash (high) answers first: 14.3 s against 247.0 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, Gemini 3.5 Flash or GPT-6 Astra?
- Gemini 3.5 Flash streams faster: 222 against 58 output tokens per second.
- Do independent benchmarks agree on Gemini 3.5 Flash vs GPT-6 Astra?
- Across the 64 benchmarks both have, Gemini 3.5 Flash leads 8, GPT-6 Astra leads 56. Independent sources split: Gemini 3.5 Flash leads on Arena Text; GPT-6 Astra leads on BenchLeader Index and AA Intelligence Index v4.3.2.
- When was GPT-6 Astra released?
- GPT-6 Astra was released on 2 Sept 2026.