Compare models
HEAD TO HEAD

GPT-6 Astra vs Muse Spark 1.3

Published test results, prices, limits and features, side by side. Share the link to show someone exactly this comparison.

At a glance

  • Capability index

    • GPT-6 Astra166.4
    • Muse Spark 1.3156.8

    GPT-6 Astra is 9.7 points higher.

  • Price per 1M tokens

    • GPT-6 Astra$20.00
    • Muse Spark 1.3$2.00

    Muse Spark 1.3 is 10 times cheaper.

  • Context window

    • GPT-6 Astra1.1M tokens
    • Muse Spark 1.31M tokens

    GPT-6 Astra takes in 0% more.

  • Released

    • GPT-6 AstraSep 4, 2026
    • Muse Spark 1.3Sep 2, 2026

    GPT-6 Astra is 2 days newer.

Where each scores higher

On the 19 tests both have taken, each on its own scale. A gap of a point or two can sit within a test’s margin of error.

GPT-6 Astra

  • LiveBench overall+0.6 points
  • Arena WebDev+131 points
  • Arena coding+5 points
  • LiveBench mathematics+0.9 points
  • Mock AIME 2024–2025+0.8 points
  • FrontierMath Tiers 1–3+19.3 points
  • FrontierMath Tier 4+51.3 points
  • LiveBench language+6.6 points
  • LiveBench data analysis+3.4 points
  • LiveBench reasoning+3.0 points
  • Chess puzzles+34.0 points

Muse Spark 1.3

  • Arena text+17 points
  • LiveBench coding+0.7 points
  • Arena math+19 points
  • LiveBench agentic coding+6.8 points
  • LiveBench instructions+2.4 points
  • Arena creative writing+10 points
  • Arena instruction following+12 points
  • Arena hard prompts+15 points

4 more tests have results for only one of them; the table below lists every result.

Everything side by side

Comparison of GPT-6 Astra vs Muse Spark 1.3
MeasureGPT-6 AstraOpenAIMuse Spark 1.3Meta AI
Model
MakerOpenAIMeta AI
Sold here byOpenAIDevPass (LLM Gateway)
API model IDgpt-6-astramuse-spark-1.3
ReleasedSep 4, 2026Sep 2, 2026
Knowledge cutoffApr 30, 2026Not reported
WeightsClosed: hosted access onlyClosed: hosted access only
Developer’s countryUnited StatesUnited States
Hosts selling it29including OpenAI directly11
Training compute1.0 × 10²⁷ FLOPEpoch AI estimateNot published
Test results
Capability index166.4 (best of these models)#2 of 274156.8#19 of 274
LiveBench overall82.2% (best of these models)#4 of 6681.6%#7 of 66
Arena text1477#29 of 4131494 (best of these models)#9 of 413
LiveBench coding80.4%#20 of 6681.1% (best of these models)#17 of 66
Arena WebDev1788 (best of these models)#2 of 1381657#14 of 138
Arena coding1543 (best of these models)#7 of 4081539#11 of 408
LiveBench mathematics96.8% (best of these models)#4 of 6696.0%#12 of 66
Mock AIME 2024–2025100.0% (best of these models)#1 of 29799.2%#15 of 297
FrontierMath Tiers 1–393.7% (best of these models)#1 of 11474.4%#19 of 114
FrontierMath Tier 497.6% (best of these models)#2 of 7046.3%#25 of 70
Arena math1490#27 of 3961509 (best of these models)#9 of 396
SimpleQA Verified75.6%#1 of 86Not tested
LiveBench agentic coding57.3%#21 of 6664.1% (best of these models)#9 of 66
DeepSWE74.1%#1 of 69Not tested
CursorBenchNot tested41.6%#22 of 62
LiveBench language89.4% (best of these models)#4 of 6682.8%#24 of 66
LiveBench instructions75.6%#8 of 6678.0% (best of these models)#4 of 66
Arena creative writing1449#39 of 4111459 (best of these models)#28 of 411
Arena instruction following1474#30 of 4131486 (best of these models)#16 of 413
LiveBench data analysis83.0% (best of these models)#1 of 6679.6%#12 of 66
LiveBench reasoning92.7% (best of these models)#1 of 6689.7%#15 of 66
GPQA Diamond95.8%#1 of 319Not tested
Chess puzzles72.0% (best of these models)#1 of 22738.0%#32 of 227
Arena hard prompts1502#30 of 4131517 (best of these models)#10 of 413
Price per million tokens
Input$10.00$1.25 (best of these models)
Output$50.00$4.25 (best of these models)
Blended, 3 input to 1 output$20.00$2.00 (best of these models)
Cached input$1.00$0.15 (best of these models)
Writing to the cache$12.50Not reported
Long requests$20.00 in, $75.00 outabove 272K tokensNo separate price reported
The maker’s own priceThis listingNot sold directly
Limits
Context window1.1M tokens (best of these models)1M tokens
Max input922K tokens1M tokens (best of these models)
Max output128K tokens131.1K tokens (best of these models)
Features
Readsimages, PDFs and textaudio, images, PDFs, text and video
Producestexttext
ReasoningYeseffort low, medium, high, xhigh, maxYeseffort minimal, low, medium, high, xhigh
Tool callingYesYes
Structured outputYes, such as JSONYes, such as JSON
File attachmentsYesYes
Temperature settingFixed by the providerSupported

Bold marks the better value in each row. Prices are each model’s own list price, or the middle price across hosts when the maker does not sell it, blended as three input tokens for every output token. Test results as published by LiveBench (Apache-2.0), Epoch AI (CC BY 4.0), Datacurve DeepSWE, compiled by Epoch AI (CC BY 4.0), Cursor CursorBench, compiled by Epoch AI (CC BY 4.0), Arena (CC BY 4.0); model details and prices from models.dev (MIT). Confirm pricing with the provider before use.