GPT-6.1 Terra vs GPT-6.2 Astra Pro
Full profiles: GPT-6.1 Terra (OpenAI) and GPT-6.2 Astra Pro (OpenAI): every benchmark, price and rank.
Updated · Data from 11 sources
As of 21 Nov 2026, there are not enough shared benchmarks for a verdict, and GPT-6.1 Terra is 8.0× cheaper at $2.00 vs $16.00 per 1M tokens blended. GPT-6.2 Astra Pro leads in every category we measure both in.
GPT-6.2 Astra Pro was released on 17 Nov 2026 and has no BenchLeader Index yet (it needs five independent results across two categories). So this page gives no quality verdict: it compares what is published so far for both, price, context window, speed and the benchmarks both have results on. First results are tracked on GPT-6.2 Astra Pro's page.
Key facts (as of 21 Nov 2026)
- GPT-6.1 Terra costs $2.00 per 1M tokens blended (list price, 21 Nov 2026).
- GPT-6.2 Astra Pro costs $16.00 per 1M tokens blended (list price, 21 Nov 2026).
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 |
|---|---|---|---|
| Reasoning | GPT-6.2 Astra Pro | 68 vs 67 category score | Medium |
| Chat quality (human preference) | – | Only GPT-6.1 Terra has an Arena rating (66 category score) | Not comparable |
| Lowest cost | GPT-6.1 Terra | $2.00 vs $16.00 per 1M blended; agentic coding session $0.080 vs $0.640 | – |
| Fastest first answer | GPT-6.2 Astra Pro | 8.3 s (GPT-6.2 Astra Pro (max)) vs 167.7 s (GPT-6.1 Terra (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
- Not enough shared benchmarks for a verdict.
- GPT-6.2 Astra Pro (max) is stronger in reasoning.
- GPT-6.1 Terra (max) is 8.0× cheaper ($2.00 vs $16.00 per 1M blended).
At a glance
| Spec | GPT-6.1 Terra | GPT-6.2 Astra Pro |
|---|---|---|
| Developer | OpenAI | OpenAI |
| Released | 26 Oct 2026 | 17 Nov 2026 |
| Context window (tokens) | 1,050,000 | 1,050,000 |
| Compared at | GPT-6.1 Terra (max) | GPT-6.2 Astra Pro (max) |
| Open weights | No | No |
| Price in / out / cached input per 1M | $1.00 / $5.00 / $0.050 | $8.00 / $40.00 / $0.400 |
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 | GPT-6.1 Terra (max) | GPT-6.2 Astra Pro (max) | Leader |
|---|---|---|---|
| Reasoning | 67.1 | 68.1 | GPT-6.2 Astra Pro |
| Coding | 68.2 | – | not comparable |
| Agents & tools | 61.6 | – | not comparable |
| Maths | 69.3 | – | not comparable |
| Knowledge | 60.6 | – | not comparable |
| Human preference | 66.2 | – | not comparable |
| Multimodal | 62.4 | – | not comparable |
| Long context | 66.1 | – | not comparable |
| Composite | 62.6 | – | not comparable |
On one chart
Each domain on the index's own scale, where 50 is an average model. Click a domain to see the benchmarks behind it. Pick up to 5 models from the list.
Dashed: that model has no result on the axis it crosses; the line only joins its neighbours.
Not drawn, no result for these models: Instruction following.
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. Cost, tokens and time per benchmark task are measured: Artificial Analysis’ averages over its whole Intelligence Index run.
| Measure | GPT-6.1 Terra (max) | GPT-6.2 Astra Pro (max) |
|---|---|---|
| Blended price per 1M tokens | $2.00 | $16.00 |
| Input / output per 1M | $1.00 / $5.00 | $8.00 / $40.00 |
| Cached input per 1M | $0.050 | $0.400 |
| Price per task: chat reply | $0.0019 | $0.015 |
| Price per task: summarize a 30-page report | $0.015 | $0.120 |
| Price per task: code edit | $0.013 | $0.108 |
| Price per task: agentic coding session | $0.080 | $0.640 |
| Price per task: structured extraction | $0.0030 | $0.024 |
| Agentic coding session, input mostly cached | $0.037 | $0.298 |
| Cost per benchmark task (AA Index) | $0.961 | – |
| Output tokens per task (AA Index) | 60k | – |
| Of which reasoning | 39k | – |
| Time per task (AA Index) | 16 min | – |
| Output speed | 63 tok/s | 56 tok/s |
| Time to first answer | 167.7 s | 8.3 s |
| Time to a complete answer, last measured 13 Nov 2026 | 375.7 s | – |
Each model is compared at its best-ranked reasoning-effort setting. GPT-6.1 Terra (max) scores 63.1 and takes 168 s to its first answer token; GPT-6.1 Terra (high) scores 59.8 in 12 s, GPT-6.1 Terra (xhigh) scores 59.5 in 34 s. GPT-6.2 Astra Pro (max) and takes 8 s to its first answer token. 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
94 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 | GPT-6.1 Terra (max) | GPT-6.2 Astra Pro (max) |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 63.1 | – |
| Agents & tools score | BenchLeader | 61.6 | – |
| Coding score | BenchLeader | 68.2 | – |
| Composite score | BenchLeader | 62.6 | – |
| Human preference score | BenchLeader | 66.2 | – |
| Knowledge score | BenchLeader | 60.6 | – |
| Long context score | BenchLeader | 66.1 | – |
| Maths score | BenchLeader | 69.3 | – |
| Multimodal score | BenchLeader | 62.4 | – |
| Reasoning score | BenchLeader | 67.1 | 68.1 |
| Blended price $/M | BenchLeader | $2.00 | $16.00 |
| Output speed | BenchLeader | 63 tok/s | 56 tok/s |
| Time to first answer | BenchLeader | 167.7 s | 8.3 s |
| Context window | BenchLeader | 1,050,000 tokens | 1,050,000 tokens |
| Reasoning | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | 94.3% | – |
| SimpleBench | SimpleBench | 72.2%† | 79.2% |
| Arena Hard Prompts | Arena (LMArena) | 1500 | – |
| LiveBench 2.0 Reasoning | LiveBench | 88.7% | – |
| Humanity's Last Exam (AA) | Artificial Analysis | 46.3% | – |
| HLE-Diamond (AA) | Artificial Analysis | 47.3% | – |
| CritPt | Artificial Analysis | 28.0% | – |
| CritPt v2 hard subset (AA) | Artificial Analysis | 68.0% | – |
| HLE-Diamond (Scale) | Scale AI SEAL | 35.7%† | – |
| Arena Expert | Arena (LMArena) | 1516 | – |
| Mystery Game Puzzles | Epoch AI Benchmarking Hub | 74.0% | – |
| Coding | |||
| SciCode | SciCode | 55.1% | – |
| Arena Coding | Arena (LMArena) | 1541 | – |
| Arena WebDev | Arena (LMArena) | 1685 | – |
| LiveBench 2.0 Coding | LiveBench | 89.2% | – |
| Vibe Code Bench v1.1 | Vals AI | 84.0% | – |
| SWE Atlas v1.1 | Scale AI SEAL | 57.7% | – |
| SWE Atlas v1.1: Codebase QnA | Scale AI SEAL | 54.0% | – |
| SWE Atlas v1.1: Test Writing | Scale AI SEAL | 63.3% | – |
| SWE Atlas v1.1: Refactoring | Scale AI SEAL | 55.7% | – |
| MirrorCode | Epoch AI Benchmarking Hub | 76.7% | – |
| ALE-Bench | Epoch AI Benchmarking Hub | 2608.8 | – |
| WeirdML v3 | WeirdML | 28.1%† | – |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 55.6% | – |
| GameBench 2 Overall | DuelLab GameBench 2 | 81.2† | – |
Show all 94 benchmarks
| Metric | Publisher | GPT-6.1 Terra (max) | GPT-6.2 Astra Pro (max) |
|---|---|---|---|
| Coding | |||
| GameBench 2 by reasoning setting | DuelLab GameBench 2 | 85.8† | – |
| Design Arena Website | Design Arena | 1287† | – |
| Design Arena Code | Design Arena | 1293† | – |
| FrontierCode 1.1 (AA) | Artificial Analysis | 36.0% | – |
| FrontierCode 2.0 (AA) | Artificial Analysis | 42.3% | – |
| Agents & tools | |||
| Terminal-Bench | Terminal-Bench | 53.3% | – |
| APEX-Agents | Mercor | 68.3% | – |
| Arena Agent | Arena (LMArena) | 10.5 | – |
| LiveBench Long Agentic | LiveBench | 55.3% | – |
| LiveBench Browser | LiveBench | 93.1% | – |
| HiL-Bench | Scale AI SEAL | 68.0%† | – |
| GDPval-AA v2.1 | Artificial Analysis | 1476 | – |
| Finance Agent v2.1 | Vals AI | 56.0% | – |
| Harvey's Legal Agent Benchmark v1.1 | Vals AI | 1.7% | – |
| Legal Research Bench | Vals AI | 26.0% | – |
| Tax Agent Bench | Vals AI | 17.1% | – |
| Terminal-Bench 4.0 (Vals) | Vals AI | 49.5% | – |
| Arena Agent Task Completion | Arena (LMArena) | 8.5 | – |
| Vending-Bench 2 | Andon Labs | 9107.5† | – |
| SpireClimb | Epoch AI Benchmarking Hub | 22.3% | – |
| AutomationBench | Artificial Analysis | 60.2% | – |
| GDP.pdf (AA) | Artificial Analysis | 28.6% | – |
| AA-Briefcase v1.1 | Artificial Analysis | 1399 | – |
| GDPx | GDPx by SemiAnalysis | 8.5% | – |
| Design Arena Agentic | Design Arena | 1174† | – |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | 57.6% | – |
| MKA (AA) | Artificial Analysis | 43.8% | – |
| FACTS Search | 92.7%† | – | |
| Maths | |||
| FrontierMath Tiers 1–3 | Epoch AI Benchmarking Hub | 90.9% | – |
| FrontierMath Tier 4 | Epoch AI Benchmarking Hub | 92.7% | – |
| OTIS Mock AIME | Epoch AI Benchmarking Hub | 100.0% | – |
| Arena Maths | Arena (LMArena) | 1479 | – |
| Knowledge | |||
| SimpleQA Verified | Epoch AI Benchmarking Hub | 57.3% | – |
| LiveBench 2.0 Data Analysis | LiveBench | 82.3% | – |
| LiveBench Research | LiveBench | 79.1% | – |
| AA-Omniscience v1.1 | Artificial Analysis | 29.6 | – |
| Excel Modeling Benchmark v1.1 | Vals AI | 73.3% | – |
| DrugDiscoveryBench | Scale AI SEAL | 58.1%† | – |
| Arena Text Factuality | Arena (LMArena) | 1455 | – |
| EBR-bench | Epoch AI Benchmarking Hub | 78.1% | – |
| AA-Omniscience v1.1: accuracy | Artificial Analysis | 47.6% | – |
| AA-Omniscience v1.1: non-hallucination | Artificial Analysis | 65.5% | – |
| FACTS Parametric | 70.8%† | – | |
| Instruction following | |||
| MultiChallenge | Scale AI SEAL | 74.2%† | – |
| Arena Instruction Following | Arena (LMArena) | 1468 | – |
| Human preference | |||
| Arena Text | Arena (LMArena) | 1474 | – |
| Arena Creative Writing | Arena (LMArena) | 1449 | – |
| Arena Multi-turn | Arena (LMArena) | 1477 | – |
| Arena Longer Queries | Arena (LMArena) | 1485 | – |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | 1247 | – |
| LiveBench Design | LiveBench | 79.7% | – |
| MMMU-Pro | Artificial Analysis | 84.7% | – |
| Blueprint-Bench 2 | Epoch AI Benchmarking Hub | 42.1%† | – |
| Furniture Assembly | Epoch AI Benchmarking Hub | 79.2% | – |
| VIALS-AA | Handshake AI | 60.3% | – |
| FACTS Multimodal | 48.8%† | – | |
| Long context | |||
| AA-LCR v1.1 | Artificial Analysis | 82.7% | – |
| MLCR | Artificial Analysis | 30.6% | – |
| FACTS Grounding | 84.2%† | – | |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | 167.7† | – |
| LiveBench 2.0 | LiveBench | 78.6% | – |
| AA Intelligence Index v5.0.1 | Artificial Analysis | 51.3 | – |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 48.1 | – |
| AA Progress Index v1 | Artificial Analysis | 345.2 | – |
| Vals Index | Vals AI | 60.5 | – |
| Vals Finance Index | Vals AI | 53.3% | – |
| Vals Legal Index | Vals AI | 58.8% | – |
| Safety & honesty | |||
| FORTRESS | Scale AI SEAL | 18.2%† | – |
| DistressBench | Scale AI SEAL | 65.9%† | – |
Data as of 2026-11-21. Best configuration of each model; every score links to its source on the model pages.
DuelLab GameBench 2, release gb2-20261119-1d7fd529faf6 (https://benchmarks.duellab.org/releases/gb2-20261119-1d7fd529faf6/), CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Scores mapped to BenchLeader model names; not endorsed by DuelLab.
What changed
- GPT-6.1 Terra (max): 4 new results (SWE Atlas v1.1, SWE Atlas v1.1: Codebase QnA, SWE Atlas v1.1: Test Writing…).
- GPT-6.1 Terra (max): 11 new results (Arena Text, Arena Hard Prompts, Arena Coding…).
- GPT-6.1 Terra (max): 3 score revisions.
- GPT-6.1 Terra (max): BenchLeader Index 64.1 → 63.5.
- GPT-6.1 Terra (max): 2 score revisions.
- GPT-6.1 Terra (max): 8 new results (LiveBench 2.0, LiveBench Long Agentic, LiveBench 2.0 Reasoning…).
- GPT-6.1 Terra (max): BenchLeader Index 64.7 → 64.1.
- GPT-6.1 Terra (max): 1 new result (Terminal-Bench).
- GPT-6.1 Terra (max): BenchLeader Index 65.2 → 64.7.
- GPT-6.1 Terra (max): 1 new result (MirrorCode).
Related comparisons
- GPT-6.1 Terra vs Claude Fable 5.2
- GPT-6.1 Terra vs Gemini 4 Carbon
- GPT-6.1 Terra vs GPT-6.2 Astra
- GPT-6.1 Terra vs Gemini 4 Neon
- GPT-6.1 Terra vs GPT-6.1 Sol
- GPT-6.1 Terra vs GPT-6 Astra
- GPT-6.1 Terra vs Claude Opus 5.5
- GPT-6.1 Terra vs Claude Fable 5.1
Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs
GPT-6.1 Terra vs GPT-6.2 Astra Pro: questions
- GPT-6.1 Terra vs GPT-6.2 Astra Pro: which is cheaper, and will that change?
- GPT-6.1 Terra is cheaper today: $2.00 against $16.00 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
- Which is faster, GPT-6.1 Terra or GPT-6.2 Astra Pro?
- GPT-6.1 Terra streams faster (63 vs 56 output tokens per second, Artificial Analysis 7-day median).
- Which is better for long documents?
- Both accept 1,050,000 tokens of context.
- Which has lower latency?
- GPT-6.2 Astra Pro answers first: 8.3 s against 167.7 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.
- When was GPT-6.2 Astra Pro released?
- GPT-6.2 Astra Pro was released on 17 Nov 2026.