Gemini 4 Carbon vs GPT-6 Astra
Gemini 4 Carbon leads GPT-6 Astra on coding, 86.6 to 83.8 on the Vibe Coding Index, and Gemini 4 Carbon costs 5.0x less per output token ($10.00 vs $50.00 per million). Updated November 14, 2026; 72 benchmarks have a result for both.
vs boards won of 72 shared, 2 tied
Gemini 4 Carbon or GPT-6 Astra: which to pick
Pick Gemini 4 Carbon for coding: it scores 86.6 on the Vibe Coding Index against 83.8 for GPT-6 Astra, though on 4 benchmarks against 5.
Pick GPT-6 Astra for a larger context window (1.05M against 1.05M tokens) and work measured by Vibe Code Bench (Vals AI) or FrontierCode 2.0, where it leads.
Model AGemini 4 CarbonGoogle86.6/1004 independent benchmarksNo cost per solved taskModel BGPT-6 AstraOpenAI83.8/1005 independent benchmarksNo cost per solved task
- Shared results
- 72
- Gemini 4 Carbon only
- 1
- GPT-6 Astra only
- 74
- Like-for-like categories
- 8 of 9
Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.
Which one for your work
Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.
Coding
Vibe Coding Index, six public coding benchmarks
Gemini 4 Carbon
Gemini 4 Carbon scores 86.6 against 83.8.
4 and 5 benchmarks behind the scores
Intelligence
AA Intelligence Index
Gemini 4 Carbon
Gemini 4 Carbon leads, 60.0 against 56.9.
Same board, both measured
Agentic work
AA Agentic Index
No call
Only GPT-6 Astra has a AA Agentic Index result, 51.0.
Directional only
Web design
Design Arena: websites
GPT-6 Astra
GPT-6 Astra leads, 1340 against 1319.
Same board, both measured
General
Vals Index
Gemini 4 Carbon
Gemini 4 Carbon leads, 70.5% against 65.0%.
Same board, both measured
Long documents
Listed context window
GPT-6 Astra
GPT-6 Astra takes 1.05M tokens against 1.05M.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Gemini 4 Carbon
Gemini 4 Carbon $0.0070, GPT-6 Astra $0.035 per turn.
List prices
At a glance
- Coding index
- Gemini 4 Carbon 86.6 from 4 benchmarks, GPT-6 Astra 83.8 from 5 benchmarks.
- Widest category
- General, 13.2 points apart on average over 16 shared benchmarks.
- Widest single gap
- Tax Agent Bench (Vals AI): Gemini 4 Carbon 51.8% against GPT-6 Astra 20.7%, 31.1 points apart on a 0 to 100 scale.
- Output price
- Gemini 4 Carbon $10.00 against GPT-6 Astra $50.00 per million output tokens, 5.0x apart.
- Context window
- Gemini 4 Carbon takes 1.05M tokens, GPT-6 Astra 1.05M.
Benchmark by benchmark
First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.
| Category | Gemini 4 Carbon | Bars | GPT-6 Astra | Ties |
|---|---|---|---|---|
| Coding | 7 | 5 | ||
| Intelligence | 2 | 0 | ||
| Agentic work | 7 | 3 | ||
| Reasoning | 8 | 4 | 1 tie | |
| Math | 2 | 2 | 1 tie | |
| Knowledge | 8 | 0 | ||
| Web design | 1 | 5 | ||
| General | 14 | 2 | ||
| All shared boards | 49 | 21 | 2 ties |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index75
| Benchmark | Gemini 4 Carbon | GPT-6 Astra | Leader |
|---|---|---|---|
| LiveBench coding | 91.0% | 87.5% | Gemini 4 Carbon |
| LMArena WebDev | 1768 | 1764 | Gemini 4 Carbon |
| AA Coding Index | Not measured | 76.9 | Only GPT-6 Astra measured |
| Terminal-Bench 4 (Vals AI) | 63.1% | 59.6% | Gemini 4 Carbon |
| Vibe Code Bench (Vals AI) | 89.1% | 89.6% | GPT-6 Astra |
| Code migration (Vals AI) | Not measured | 67.7% | Only GPT-6 Astra measured |
| ProgramBench (Vals AI) | Not measured | 5.5% | Only GPT-6 Astra measured |
| IOI (Vals AI) | Not measured | 100.0% | Only GPT-6 Astra measured |
| Terminal-Bench 2.1 (Vals AI) | Not measured | 87.3% | Only GPT-6 Astra measured |
| SWE-bench Multimodal (Vals AI) | Not measured | 43.3% | Only GPT-6 Astra measured |
| Vibe Code Bench 1-100 (Vals AI) | Not measured | 27.6% | Only GPT-6 Astra measured |
| CyberBench (Vals AI) | Not measured | 41.1% | Only GPT-6 Astra measured |
| SciCode | Not measured | 56.5% | Only GPT-6 Astra measured |
| Terminal-Bench 4.0 (Artificial Analysis run) | 65.2% | 59.6% | Gemini 4 Carbon |
| FrontierCode 2.0 | 40.6% | 42.4% | GPT-6 Astra |
| FrontierCode | Not measured | 53.3% | Only GPT-6 Astra measured |
| FrontierSWE | 61.2% | 65.5% | GPT-6 Astra |
| GSO-Bench | Not measured | 79.4% | Only GPT-6 Astra measured |
| WeirdML | Not measured | 93.3% | Only GPT-6 Astra measured |
| MirrorCode | Not measured | 60.0% | Only GPT-6 Astra measured |
| ALE-Bench | 2374.7 | 2951.3 | GPT-6 Astra |
| LMArena Text: coding | 1564 | 1494 | Gemini 4 Carbon |
| LMArena Text: software and IT | 1552 | 1479 | Gemini 4 Carbon |
| LMArena WebDev: React | 1792 | 1776 | Gemini 4 Carbon |
| LMArena WebDev: HTML | 1770 | 1786 | GPT-6 Astra |
| LMArena WebDev: from an image | Not measured | 1722 | Only GPT-6 Astra measured |
| SciCode | Not measured | 12.3% | Only GPT-6 Astra measured |
| SWE-Bench Pro V2 | Not measured | 96.9% | Only GPT-6 Astra measured |
| SWE Atlas: Codebase QnA | Not measured | 59.1% | Only GPT-6 Astra measured |
| SWE Atlas: Refactoring | Not measured | 59.1% | Only GPT-6 Astra measured |
| SWE Atlas: Test Writing | Not measured | 50.7% | Only GPT-6 Astra measured |
| GameBench 2 | Not measured | 94.3 | Only GPT-6 Astra measured |
| Terminal-Bench 2.1 | Not measured | 87.4% | Only GPT-6 Astra measured |
| Terminal-Bench 4.0 | Not measured | 58.2% | Only GPT-6 Astra measured |
| CursorBench 4.0 | 60.4% | Not measured | Only Gemini 4 Carbon measured |
| ProgramBench | Not measured | 5.0% | Only GPT-6 Astra measured |
| Neither model has a result on the other 16 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | Gemini 4 Carbon | GPT-6 Astra | Leader |
|---|---|---|---|
| AA Intelligence Index | 60.0 | 56.9 | Gemini 4 Carbon |
| Epoch Capabilities Index | 169.8 | 169.0 | Gemini 4 Carbon |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index73
| Benchmark | Gemini 4 Carbon | GPT-6 Astra | Leader |
|---|---|---|---|
| AA Agentic Index | Not measured | 51.0 | Only GPT-6 Astra measured |
| Design Arena: agents building web apps | Not measured | 1359 | Only GPT-6 Astra measured |
| Design Arena: agents building full-stack apps | Not measured | 1333 | Only GPT-6 Astra measured |
| Design Arena: agents building mobile apps | 1284 | 1280 | Gemini 4 Carbon |
| Design Arena: agents building Android apps | Not measured | 1186 | Only GPT-6 Astra measured |
| Design Arena: agents building Godot games | Not measured | 1319 | Only GPT-6 Astra measured |
| Design Arena: HTML slides | Not measured | 1530 | Only GPT-6 Astra measured |
| Design Arena: PowerPoint slides | Not measured | 1369 | Only GPT-6 Astra measured |
| Design Arena: spreadsheets | 1224 | 1332 | GPT-6 Astra |
| LiveBench research | 92.5% | 90.6% | Gemini 4 Carbon |
| LiveBench browser | 80.0% | 91.3% | GPT-6 Astra |
| LiveBench long agentic | 72.7% | 75.6% | GPT-6 Astra |
| SREBench (Vals AI) | Not measured | 56.9% | Only GPT-6 Astra measured |
| Finance Agent (Vals AI) | Not measured | 53.5% | Only GPT-6 Astra measured |
| Tax Agent Bench (Vals AI) | 51.8% | 20.7% | Gemini 4 Carbon |
| Legal Research Bench (Vals AI) | 55.8% | 39.4% | Gemini 4 Carbon |
| Harvey's Legal Agent Benchmark (Vals AI) | 17.5% | 7.1% | Gemini 4 Carbon |
| Excel Modeling Benchmark (Vals AI) | Not measured | 71.7% | Only GPT-6 Astra measured |
| BioMysteryBench (Vals AI) | Not measured | 79.3% | Only GPT-6 Astra measured |
| Terminal-Bench-Science (Vals AI) | Not measured | 62.9% | Only GPT-6 Astra measured |
| Vals Science Index | Not measured | 65.1% | Only GPT-6 Astra measured |
| Time Horizon Index (Vals AI) | Not measured | 90.5% | Only GPT-6 Astra measured |
| CUA-bench (Vals AI) | Not measured | 19.2% | Only GPT-6 Astra measured |
| Robo-bench (Vals AI) | Not measured | 23.3% | Only GPT-6 Astra measured |
| APEX-Agents | 84.2% | 64.7% | Gemini 4 Carbon |
| Remote Labor Index | Not measured | 20.8% | Only GPT-6 Astra measured |
| BALROG | Not measured | 68.3% | Only GPT-6 Astra measured |
| PostTrainBench | Not measured | 41.9% | Only GPT-6 Astra measured |
| SpireClimb | Not measured | 34.0% | Only GPT-6 Astra measured |
| Vending-Bench 2 | $16,484 | $15,515 | Gemini 4 Carbon |
| LMArena Search | Not measured | 1231 | Only GPT-6 Astra measured |
| Neither model has a result on the other 13 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index84
| Benchmark | Gemini 4 Carbon | GPT-6 Astra | Leader |
|---|---|---|---|
| LiveBench reasoning | 90.6% | 92.7% | GPT-6 Astra |
| MysteryMechanism (Vals AI) | Not measured | 53.1% | Only GPT-6 Astra measured |
| GPQA Diamond (Artificial Analysis run) | Not measured | 96.3% | Only GPT-6 Astra measured |
| Humanity's Last Exam (Artificial Analysis run) | Not measured | 54.7% | Only GPT-6 Astra measured |
| HLE-Diamond (Artificial Analysis run) | 69.1% | 63.8% | Gemini 4 Carbon |
| CritPt v2, hard subset (Artificial Analysis run) | 69.0% | 72.0% | GPT-6 Astra |
| GPQA Diamond (Epoch AI run) | 96.5% | 95.8% | Gemini 4 Carbon |
| ARC-AGI | 98.5% | 98.5% | Tie |
| ARC-AGI-2 | 94.2% | 95.0% | GPT-6 Astra |
| SimpleBench | 92.6% | 83.6% | Gemini 4 Carbon |
| Chess puzzles | 78.0% | 74.0% | Gemini 4 Carbon |
| Mystery game puzzles | 73.0% | 84.0% | GPT-6 Astra |
| EBR-bench | 87.6% | 76.2% | Gemini 4 Carbon |
| LMArena Text: hard prompts | 1551 | 1462 | Gemini 4 Carbon |
| LMArena Text: expert | 1582 | 1493 | Gemini 4 Carbon |
| LMArena Text: hard prompts in English | 1545 | 1468 | Gemini 4 Carbon |
| LMArena Vision: diagrams | Not measured | 1299 | Only GPT-6 Astra measured |
| LMArena Vision: homework | Not measured | 1332 | Only GPT-6 Astra measured |
| Humanity's Last Exam (Scale run) | Not measured | 60.6% | Only GPT-6 Astra measured |
| GoBench | Not measured | 2566 | Only GPT-6 Astra measured |
| Neither model has a result on the other 6 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index22
| Benchmark | Gemini 4 Carbon | GPT-6 Astra | Leader |
|---|---|---|---|
| OTIS Mock AIME | 100.0% | 100.0% | Tie |
| FrontierMath, tiers 1 to 3 | 90.5% | 93.7% | GPT-6 Astra |
| FrontierMath, tier 4 | 95.1% | 97.6% | GPT-6 Astra |
| ProofBench | Not measured | 99.0% | Only GPT-6 Astra measured |
| LMArena Text: math | Not measured | 1459 | Only GPT-6 Astra measured |
| LMArena Text: mathematical work | 1538 | 1478 | Gemini 4 Carbon |
| MathArena | 85.4% | 78.0% | Gemini 4 Carbon |
| Neither model has a result on the other 7 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index80
| Benchmark | Gemini 4 Carbon | GPT-6 Astra | Leader |
|---|---|---|---|
| MedCode (Vals AI) | Not measured | 48.5% | Only GPT-6 Astra measured |
| AA-Omniscience | 53.5 | 44.6 | Gemini 4 Carbon |
| SimpleQA Verified | 77.7% | 75.6% | Gemini 4 Carbon |
| LMArena Text: science | 1535 | 1447 | Gemini 4 Carbon |
| LMArena Text: medicine and healthcare | 1525 | 1420 | Gemini 4 Carbon |
| LMArena Text: legal and government | 1542 | 1453 | Gemini 4 Carbon |
| LMArena Text: business and finance | 1520 | 1442 | Gemini 4 Carbon |
| FACTS Benchmark Suite | 78.4% | 78.1% | Gemini 4 Carbon |
| PRBench Finance | Not measured | 47.5% | Only GPT-6 Astra measured |
| PRBench Legal | 53.3% | 48.4% | Gemini 4 Carbon |
| Neither model has a result on the other 11 boards here | |||
Web designLike-for-likePublic boards, shown as published, kept out of the index15
| Benchmark | Gemini 4 Carbon | GPT-6 Astra | Leader |
|---|---|---|---|
| Design Arena: websites | 1319 | 1340 | GPT-6 Astra |
| Design Arena: UI components | 1357 | 1401 | GPT-6 Astra |
| Design Arena: game dev | 1326 | 1468 | GPT-6 Astra |
| Design Arena: data viz | Not measured | 1335 | Only GPT-6 Astra measured |
| Design Arena: 3D | 1434 | 1524 | GPT-6 Astra |
| Design Arena: all code categories | 1348 | 1395 | GPT-6 Astra |
| Design Arena: SVG | Not measured | 1421 | Only GPT-6 Astra measured |
| Design Arena: ASCII art | Not measured | 1343 | Only GPT-6 Astra measured |
| LiveBench design | 80.2% | 80.0% | Gemini 4 Carbon |
GeneralLike-for-likePublic boards, shown as published, kept out of the index142
| Benchmark | Gemini 4 Carbon | GPT-6 Astra | Leader |
|---|---|---|---|
| Vals Index | 70.5% | 65.0% | Gemini 4 Carbon |
| LiveBench data analysis | 86.5% | 83.0% | Gemini 4 Carbon |
| MedScribe (Vals AI) | Not measured | 87.9% | Only GPT-6 Astra measured |
| Long context reasoning (AA-LCR) | 79.3% | 80.7% | GPT-6 Astra |
| GDPval-AA | 1733 | 1587 | Gemini 4 Carbon |
| DTBench | Not measured | 97.3% | Only GPT-6 Astra measured |
| LMCA | Not measured | 64.4% | Only GPT-6 Astra measured |
| Furniture assembly | 77.5% | 80.0% | GPT-6 Astra |
| LMArena Text | 1538 | 1440 | Gemini 4 Carbon |
| LMArena Text: instruction following | 1539 | 1456 | Gemini 4 Carbon |
| LMArena Text: creative writing | 1533 | 1421 | Gemini 4 Carbon |
| LMArena Vision | Not measured | 1279 | Only GPT-6 Astra measured |
| LMArena Text: English | 1530 | 1448 | Gemini 4 Carbon |
| LMArena Text: non-English | 1534 | 1430 | Gemini 4 Carbon |
| LMArena Text: Chinese | 1613 | 1496 | Gemini 4 Carbon |
| LMArena Text: German | Not measured | 1465 | Only GPT-6 Astra measured |
| LMArena Text: Japanese | Not measured | 1415 | Only GPT-6 Astra measured |
| LMArena Text: Korean | Not measured | 1424 | Only GPT-6 Astra measured |
| LMArena Text: French | Not measured | 1440 | Only GPT-6 Astra measured |
| LMArena Text: Spanish | Not measured | 1436 | Only GPT-6 Astra measured |
| LMArena Text: Russian | 1560 | 1441 | Gemini 4 Carbon |
| LMArena Text: Polish | Not measured | 1426 | Only GPT-6 Astra measured |
| LMArena Text: multi-turn | 1569 | 1448 | Gemini 4 Carbon |
| LMArena Text: longer queries | 1549 | 1459 | Gemini 4 Carbon |
| LMArena Text: writing and language | 1542 | 1438 | Gemini 4 Carbon |
| LMArena Text: entertainment and media | 1527 | 1425 | Gemini 4 Carbon |
| LMArena Vision: English | Not measured | 1282 | Only GPT-6 Astra measured |
| LMArena Vision: Chinese | Not measured | 1332 | Only GPT-6 Astra measured |
| LMArena Vision: reading text | Not measured | 1290 | Only GPT-6 Astra measured |
| LMArena Vision: humor | Not measured | 1272 | Only GPT-6 Astra measured |
| LMArena Vision: creative writing | Not measured | 1278 | Only GPT-6 Astra measured |
| LMArena Document | Not measured | 1467 | Only GPT-6 Astra measured |
| Neither model has a result on the other 17 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Gemini 4 Carbon | GPT-6 Astra | Leader |
|---|---|---|---|
| Neither model has a result on 5 boards here | |||
What each workload costs
Three fixed token mixes at each model's listed price, so per-token rates become a bill.
Chat turn
1K input + 500 output tokens
Gemini 4 Carbon$0.0070Fits in one request
GPT-6 Astra$0.035Fits in one request
Gemini 4 Carbon costs less
Repository review
50K input + 3K output tokens
Gemini 4 Carbon$0.130Fits in one request
GPT-6 Astra$0.650Fits in one request
Gemini 4 Carbon costs less
Agent loop
200K input + 10K output tokens
Gemini 4 Carbon$0.500Fits in one request
GPT-6 Astra$2.50Fits in one request
Gemini 4 Carbon costs less
Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.
Specification differences
The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.
| Specification | Google Gemini 4 Carbon | OpenAI GPT-6 Astra |
|---|---|---|
| Maker | OpenAI | |
| Weights | No public weights listed | No public weights listed |
| Released | November 10, 2026 | September 4, 2026 |
| Context window | 1.05M | 1.05M |
| Input $/Mtok | $2.00 | $10.00 |
| Output $/Mtok | $10.00 | $50.00 |
| Coding index | 86.6 | 83.8 |
| Independent measurements behind it | 4 | 5 |
| Cost per solved task | Not computable | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Gemini 4 Carbon better than GPT-6 Astra for coding?
Pick Gemini 4 Carbon for coding: it scores 86.6 on the Vibe Coding Index against 83.8 for GPT-6 Astra, though on 4 benchmarks against 5. Pick GPT-6 Astra for a larger context window (1.05M against 1.05M tokens) and work measured by Vibe Code Bench (Vals AI) or FrontierCode 2.0, where it leads.
Which is cheaper, Gemini 4 Carbon or GPT-6 Astra?
Gemini 4 Carbon costs $2.00 per million input tokens and $10.00 per million output tokens; GPT-6 Astra costs $10.00 and $50.00. Cost per solved task is known for neither of them.
Which has the bigger context window, Gemini 4 Carbon or GPT-6 Astra?
GPT-6 Astra, with 1.05M tokens against 1.05M.
How many benchmarks is the Gemini 4 Carbon vs GPT-6 Astra comparison based on?
72 benchmarks have a result for both models as of November 14, 2026. Gemini 4 Carbon's coding index rests on 4 independent benchmarks and GPT-6 Astra's on 5. Where only one model has a result, the page says so instead of naming a winner.