GPT-6.2 Astra vs Muse Wonder
GPT-6.2 Astra leads Muse Wonder on coding, 87.0 to 71.8 on the Vibe Coding Index, and Muse Wonder costs 2.0x less per output token ($20.00 vs $40.00 per million). Updated November 22, 2026; 46 benchmarks have a result for both.
GPT-6.2 Astra or Muse Wonder: which to pick
Pick GPT-6.2 Astra for coding: it scores 87.0 on the Vibe Coding Index against 71.8 for Muse Wonder.
Pick Muse Wonder for output tokens 2.0x cheaper ($20.00 vs $40.00 per million) and work measured by LMArena Text: coding or LMArena Text: software and IT, where it leads.
- Shared results
- 46
- GPT-6.2 Astra only
- 30
- Muse Wonder only
- 36
- 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
GPT-6.2 Astra
GPT-6.2 Astra scores 87.0 against 71.8.
4 and 3 benchmarks behind the scores
Intelligence
AA Intelligence Index
GPT-6.2 Astra
GPT-6.2 Astra leads, 62.1 against 43.6.
Same board, both measured
Web design
Design Arena: websites
Muse Wonder
Muse Wonder leads, 1374 against 1333.
Same board, both measured
General
Vals Index
GPT-6.2 Astra
GPT-6.2 Astra leads, 67.7% against 64.9%.
Same board, both measured
Long documents
Listed context window
GPT-6.2 Astra
GPT-6.2 Astra takes 1.05M tokens against 1.05M.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Muse Wonder
GPT-6.2 Astra $0.028, Muse Wonder $0.014 per turn.
List prices
At a glance
- Coding index
- GPT-6.2 Astra 87.0 from 4 benchmarks, Muse Wonder 71.8 from 3 benchmarks.
- Widest category
- Math, 41.3 points apart on average over 1 shared benchmark.
- Widest single gap
- Terminal-Bench-Science (Vals AI): GPT-6.2 Astra 71.4% against Muse Wonder 22.9%, 48.6 points apart on a 0 to 100 scale.
- Output price
- GPT-6.2 Astra $40.00 against Muse Wonder $20.00 per million output tokens, 2.0x apart.
- Context window
- GPT-6.2 Astra takes 1.05M tokens, Muse Wonder 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 | GPT-6.2 Astra | Bars | Muse Wonder | Ties |
|---|---|---|---|---|
| Coding | 8 | 2 | ||
| Intelligence | 2 | 0 | ||
| Agentic work | 5 | 3 | ||
| Reasoning | 3 | 2 | ||
| Math | 1 | 0 | ||
| Knowledge | 1 | 3 | ||
| Web design | 3 | 2 | ||
| General | 1 | 10 | ||
| All shared boards | 24 | 22 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index82
| Benchmark | GPT-6.2 Astra | Muse Wonder | Leader |
|---|---|---|---|
| LiveBench coding | 91.9% | Not measured | Only GPT-6.2 Astra measured |
| LMArena WebDev | 1766 | 1647 | GPT-6.2 Astra |
| Terminal-Bench 4 (Vals AI) | 63.1% | 49.5% | GPT-6.2 Astra |
| Vibe Code Bench (Vals AI) | 89.9% | 83.1% | GPT-6.2 Astra |
| Code migration (Vals AI) | Not measured | 62.5% | Only Muse Wonder measured |
| IOI (Vals AI) | Not measured | 66.9% | Only Muse Wonder measured |
| SWE-bench Multimodal (Vals AI) | Not measured | 45.6% | Only Muse Wonder measured |
| Vibe Code Bench 1-100 (Vals AI) | Not measured | 26.6% | Only Muse Wonder measured |
| CyberBench (Vals AI) | Not measured | 80.4% | Only Muse Wonder measured |
| SciCode | Not measured | 60.3% | Only Muse Wonder measured |
| Terminal-Bench 4.0 (Artificial Analysis run) | 64.1% | 52.0% | GPT-6.2 Astra |
| FrontierCode 2.0 | 49.0% | 27.9% | GPT-6.2 Astra |
| MirrorCode | 67.8% | Not measured | Only GPT-6.2 Astra measured |
| LMArena Text: coding | 1450 | 1523 | Muse Wonder |
| LMArena Text: software and IT | 1443 | 1526 | Muse Wonder |
| LMArena WebDev: React | 1770 | 1655 | GPT-6.2 Astra |
| LMArena WebDev: HTML | 1854 | 1652 | GPT-6.2 Astra |
| GameBench 2 | 97.2 | 77.7 | GPT-6.2 Astra |
| Terminal-Bench 4.0 | Not measured | 47.0% | Only Muse Wonder measured |
| ProgramBench | Not measured | 5.5% | Only Muse Wonder measured |
| Neither model has a result on the other 32 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | GPT-6.2 Astra | Muse Wonder | Leader |
|---|---|---|---|
| AA Intelligence Index | 62.1 | 43.6 | GPT-6.2 Astra |
| Epoch Capabilities Index | 172.8 | 161.7 | GPT-6.2 Astra |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index53
| Benchmark | GPT-6.2 Astra | Muse Wonder | Leader |
|---|---|---|---|
| Design Arena: agents building web apps | 1334 | Not measured | Only GPT-6.2 Astra measured |
| Design Arena: agents building full-stack apps | 1351 | Not measured | Only GPT-6.2 Astra measured |
| Design Arena: agents building mobile apps | 1305 | 1239 | GPT-6.2 Astra |
| Design Arena: agents building Android apps | 1257 | Not measured | Only GPT-6.2 Astra measured |
| Design Arena: HTML slides | Not measured | 1129 | Only Muse Wonder measured |
| Design Arena: PowerPoint slides | Not measured | 1034 | Only Muse Wonder measured |
| Design Arena: spreadsheets | 1278 | 1230 | GPT-6.2 Astra |
| LiveBench research | 91.5% | Not measured | Only GPT-6.2 Astra measured |
| LiveBench browser | 93.0% | Not measured | Only GPT-6.2 Astra measured |
| LiveBench long agentic | 70.8% | Not measured | Only GPT-6.2 Astra measured |
| Finance Agent (Vals AI) | Not measured | 56.6% | Only Muse Wonder measured |
| Tax Agent Bench (Vals AI) | 21.8% | 36.3% | Muse Wonder |
| Legal Research Bench (Vals AI) | 42.8% | 50.5% | Muse Wonder |
| Harvey's Legal Agent Benchmark (Vals AI) | 16.3% | 23.8% | Muse Wonder |
| Excel Modeling Benchmark (Vals AI) | Not measured | 67.1% | Only Muse Wonder measured |
| BioMysteryBench (Vals AI) | 80.7% | 67.0% | GPT-6.2 Astra |
| Terminal-Bench-Science (Vals AI) | 71.4% | 22.9% | GPT-6.2 Astra |
| Vals Science Index | 68.4% | 39.1% | GPT-6.2 Astra |
| CUA-bench (Vals AI) | Not measured | 7.2% | Only Muse Wonder measured |
| Robo-bench (Vals AI) | 32.7% | Not measured | Only GPT-6.2 Astra measured |
| APEX-Agents | 72.2% | Not measured | Only GPT-6.2 Astra measured |
| PostTrainBench | Not measured | 24.2% | Only Muse Wonder measured |
| SpireClimb | 41.3% | Not measured | Only GPT-6.2 Astra measured |
| Vending-Bench 2 | Not measured | $9,432 | Only Muse Wonder measured |
| LMArena Search | 1191 | Not measured | Only GPT-6.2 Astra measured |
| Neither model has a result on the other 19 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index32
| Benchmark | GPT-6.2 Astra | Muse Wonder | Leader |
|---|---|---|---|
| LiveBench reasoning | 91.6% | Not measured | Only GPT-6.2 Astra measured |
| MysteryMechanism (Vals AI) | 53.1% | 27.5% | GPT-6.2 Astra |
| Humanity's Last Exam (Artificial Analysis run) | Not measured | 50.1% | Only Muse Wonder measured |
| HLE-Diamond (Artificial Analysis run) | 72.2% | 39.0% | GPT-6.2 Astra |
| CritPt v2, hard subset (Artificial Analysis run) | 71.5% | 50.5% | GPT-6.2 Astra |
| GPQA Diamond (Epoch AI run) | 96.8% | Not measured | Only GPT-6.2 Astra measured |
| ARC-AGI | 98.5% | Not measured | Only GPT-6.2 Astra measured |
| ARC-AGI-2 | 95.8% | Not measured | Only GPT-6.2 Astra measured |
| SimpleBench | 77.3% | Not measured | Only GPT-6.2 Astra measured |
| Chess puzzles | 78.0% | Not measured | Only GPT-6.2 Astra measured |
| Mystery game puzzles | 95.0% | Not measured | Only GPT-6.2 Astra measured |
| EBR-bench | 92.4% | Not measured | Only GPT-6.2 Astra measured |
| LMArena Text: hard prompts | 1425 | 1514 | Muse Wonder |
| LMArena Text: expert | Not measured | 1516 | Only Muse Wonder measured |
| LMArena Text: hard prompts in English | 1426 | 1519 | Muse Wonder |
| Humanity's Last Exam (Scale run) | 66.9% | Not measured | Only GPT-6.2 Astra measured |
| GoBench | 2493 | Not measured | Only GPT-6.2 Astra measured |
| Neither model has a result on the other 9 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | GPT-6.2 Astra | Muse Wonder | Leader |
|---|---|---|---|
| OTIS Mock AIME | 100.0% | Not measured | Only GPT-6.2 Astra measured |
| FrontierMath, tiers 1 to 3 | 92.6% | Not measured | Only GPT-6.2 Astra measured |
| FrontierMath, tier 4 | 100.0% | Not measured | Only GPT-6.2 Astra measured |
| ProofBench | Not measured | 84.0% | Only Muse Wonder measured |
| LMArena Text: math | Not measured | 1493 | Only Muse Wonder measured |
| LMArena Text: mathematical work | Not measured | 1509 | Only Muse Wonder measured |
| MathArena | 92.9% | 51.6% | GPT-6.2 Astra |
| Neither model has a result on the other 7 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index13
| Benchmark | GPT-6.2 Astra | Muse Wonder | Leader |
|---|---|---|---|
| AA-Omniscience | 50.0 | 25.1 | GPT-6.2 Astra |
| SimpleQA Verified | 78.5% | Not measured | Only GPT-6.2 Astra measured |
| LMArena Text: science | 1415 | 1507 | Muse Wonder |
| LMArena Text: medicine and healthcare | Not measured | 1500 | Only Muse Wonder measured |
| LMArena Text: legal and government | 1415 | 1514 | Muse Wonder |
| LMArena Text: business and finance | 1413 | 1490 | Muse Wonder |
| Neither model has a result on the other 15 boards here | |||
Web designLike-for-likePublic boards, shown as published, kept out of the index32
| Benchmark | GPT-6.2 Astra | Muse Wonder | Leader |
|---|---|---|---|
| Design Arena: websites | 1333 | 1374 | Muse Wonder |
| Design Arena: UI components | 1341 | 1357 | Muse Wonder |
| Design Arena: game dev | 1501 | 1394 | GPT-6.2 Astra |
| Design Arena: data viz | Not measured | 1290 | Only Muse Wonder measured |
| Design Arena: 3D | 1430 | 1420 | GPT-6.2 Astra |
| Design Arena: all code categories | 1389 | 1384 | GPT-6.2 Astra |
| Design Arena: SVG | Not measured | 1286 | Only Muse Wonder measured |
| LiveBench design | 82.8% | Not measured | Only GPT-6.2 Astra measured |
| Neither model has a result on the other 1 board here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index110
| Benchmark | GPT-6.2 Astra | Muse Wonder | Leader |
|---|---|---|---|
| Vals Index | 67.7% | 64.9% | GPT-6.2 Astra |
| LiveBench data analysis | 84.1% | Not measured | Only GPT-6.2 Astra measured |
| Public Benefits Bench (Vals AI) | Not measured | 70.2% | Only Muse Wonder measured |
| Long context reasoning (AA-LCR) | Not measured | 80.7% | Only Muse Wonder measured |
| GDPval-AA | Not measured | 1770 | Only Muse Wonder measured |
| DTBench | Not measured | 95.2% | Only Muse Wonder measured |
| LMCA | Not measured | 56.8% | Only Muse Wonder measured |
| Furniture assembly | 89.2% | Not measured | Only GPT-6.2 Astra measured |
| LMArena Text | 1414 | 1503 | Muse Wonder |
| LMArena Text: instruction following | 1423 | 1492 | Muse Wonder |
| LMArena Text: creative writing | 1417 | 1488 | Muse Wonder |
| LMArena Text: English | 1417 | 1509 | Muse Wonder |
| LMArena Text: non-English | 1404 | 1492 | Muse Wonder |
| LMArena Text: Chinese | Not measured | 1542 | Only Muse Wonder measured |
| LMArena Text: German | Not measured | 1508 | Only Muse Wonder measured |
| LMArena Text: Japanese | Not measured | 1499 | Only Muse Wonder measured |
| LMArena Text: Korean | Not measured | 1465 | Only Muse Wonder measured |
| LMArena Text: French | Not measured | 1515 | Only Muse Wonder measured |
| LMArena Text: Spanish | Not measured | 1480 | Only Muse Wonder measured |
| LMArena Text: Russian | 1437 | 1511 | Muse Wonder |
| LMArena Text: Polish | Not measured | 1535 | Only Muse Wonder measured |
| LMArena Text: multi-turn | 1425 | 1512 | Muse Wonder |
| LMArena Text: longer queries | 1431 | 1499 | Muse Wonder |
| LMArena Text: writing and language | 1440 | 1494 | Muse Wonder |
| LMArena Text: entertainment and media | 1413 | 1481 | Muse Wonder |
| MultiChallenge | 73.2% | Not measured | Only GPT-6.2 Astra measured |
| EQ-Bench 4 | Not measured | 1325 | Only Muse Wonder measured |
| Neither model has a result on the other 22 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | GPT-6.2 Astra | Muse Wonder | 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
Muse Wonder costs less
Repository review
50K input + 3K output tokens
Muse Wonder costs less
Agent loop
200K input + 10K output tokens
Muse Wonder 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 | OpenAI GPT-6.2 Astra | Meta Muse Wonder |
|---|---|---|
| Maker | OpenAI | Meta |
| Weights | No public weights listed | No public weights listed |
| Released | November 17, 2026 | October 12, 2026 |
| Context window | 1.05M | 1.05M |
| Input $/Mtok | $8.00 | $4.00 |
| Output $/Mtok | $40.00 | $20.00 |
| Coding index | 87.0 | 71.8 |
| Independent measurements behind it | 4 | 3 |
| Cost per solved task | Not computable | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is GPT-6.2 Astra better than Muse Wonder for coding?
Pick GPT-6.2 Astra for coding: it scores 87.0 on the Vibe Coding Index against 71.8 for Muse Wonder. Pick Muse Wonder for output tokens 2.0x cheaper ($20.00 vs $40.00 per million) and work measured by LMArena Text: coding or LMArena Text: software and IT, where it leads.
Which is cheaper, GPT-6.2 Astra or Muse Wonder?
GPT-6.2 Astra costs $8.00 per million input tokens and $40.00 per million output tokens; Muse Wonder costs $4.00 and $20.00. Cost per solved task is known for neither of them.
Which has the bigger context window, GPT-6.2 Astra or Muse Wonder?
GPT-6.2 Astra, with 1.05M tokens against 1.05M.
How many benchmarks is the GPT-6.2 Astra vs Muse Wonder comparison based on?
46 benchmarks have a result for both models as of November 22, 2026. GPT-6.2 Astra's coding index rests on 4 independent benchmarks and Muse Wonder's on 3. Where only one model has a result, the page says so instead of naming a winner.