Noometry

OpenAI, proprietary

GPT-6 Astra

GPT-6 Astra by OpenAI ranks 1st of 354 ranked models on the Noometry Index as of October 2026, with a score of 70.8. Its strongest category is knowledge, where it ranks 1st. API pricing starts at $10 per million input tokens and $50 per million output tokens, with a 1.05M-token context window.

Last verified

Noometry rank
#1 of 354
Index score
70.8
Evidence
Confirmed 56 results
Provider
OpenAI
Released
September 3, 2026
Weights
Proprietary
Reasoning
Yes
Context window
1.05M
Max output
128K
Input price
$10 / M
Output price
$50 / M
Blended price
$20 / M
Output speed
Not measured
Value
#206 of 219
Knowledge cutoff
April 2026
Input
text, image, pdf

Category scores

Each category score combines every public result we have in that category.

  1. Coding 73.7
  2. Agentic & Tool Use 52.9
  3. Reasoning 85.1
  4. Math 93.5
  5. Knowledge 75.3
  6. Multimodal 55.0
  7. Multilingual 53.7
  8. Instruction Following 76.3
  9. Long Context 44.5
  10. Writing & Preference 75.3
CategoryScoreRankResults
Coding73.7#29
Agentic & Tool Use52.9#34
Reasoning85.1#110
Math93.5#25
Knowledge75.3#15
Multimodal55.0#33
Multilingual53.7#611
Instruction Following76.3#441
Long Context44.5#621
Writing & Preference75.3#74

Strengths and weaknesses

Categories where GPT-6 Astra places highest and lowest among the models ranked in each, with its score against that category's median.

Strongest categories

CategoryScorevs medianRank
Reasoning85.1+61.5#1 of 350, top 1%
Knowledge75.3+38.0#1 of 314, top 1%
Coding73.7+35.0#2 of 340, top 1%

Weakest categories

CategoryScorevs medianRank
Long Context44.5+3.6#62 of 296, top 21%
Multilingual53.7+6.3#61 of 297, top 21%
Instruction Following76.3+5.0#44 of 305, top 15%

Closest competitors

The models ranked just above and below GPT-6 Astra. When scores are this close, price and speed are often the better way to choose.

ModelRankScoreBlended $/MSpeed
Claude Fable 5.1#269.0$20—Compare
Claude Opus 5.5#368.6$8—Compare
Claude Opus 5#467.8$10—Compare
Claude Fable 5#566.8$2025Compare
GPT-6.1 Sol#665.6$4—Compare
GPT-5.6 Sol#765.0$810Compare
GPT-5.5 Pro#864.3$67.50—Compare
GPT-5.5#963.4$11.2525Compare

Sponsored placements are available on pages like this one. Advertise on Noometry

Benchmark results

Every published result we track, with its source. Bold rows are the ones used for ranking; where several exist we prefer independent runs over self-reported numbers.

Coding

BenchmarkScorePositionSettingSourceDate
DeepSWE73.2%highEpoch AI
DeepSWE67%lowEpoch AI
DeepSWE73.2%maxEpoch AI
DeepSWE72.8%mediumEpoch AI
DeepSWE74.1%#2 of 29, top 7%xhighEpoch AI
FrontierCode53.3%#4 of 37, top 11%maxEpoch AI
LMArena WebDev1786#2 of 113, top 2%LMArena2026-10-08
FrontierSWE65.5%Best of 18maxEpoch AI
SciCode55.4%highEpoch AI
SciCode54.1%lowEpoch AI
SciCode56.5%#19 of 121, top 16%maxEpoch AI
SciCode54.2%mediumEpoch AI
SciCode53.5%noneEpoch AI
SciCode55.7%xhighEpoch AI
GSO79.4%#2 of 31, top 7%Epoch AI
WeirdML92.9%highEpoch AI
WeirdML93.3%maxEpoch AI
WeirdML93.6%Best of 119promaxEpoch AI
LMArena Coding1487#35 of 294, top 12%LMArena2026-10-08
MirrorCode46.7%#4 of 9, top 45%highEpoch AI2026-08-30
ALE-Bench2,951Best of 105maxEpoch AI

Agentic & Tool Use

BenchmarkScorePositionSettingSourceDate
APEX-Agents64.7%#8 of 49, top 17%Epoch AI
Remote Labor Index20.8%Best of 14Epoch AI
BALROG68.3%Best of 35maxEpoch AI
GDP.pdf34.2%Best of 36maxEpoch AI
Vending-Bench 215,515Best of 60Epoch AI

Reasoning

BenchmarkScorePositionSettingSourceDate
ARC-AGI-292.1%highEpoch AI
ARC-AGI-285.4%lowEpoch AI
ARC-AGI-295%Best of 83maxEpoch AI
ARC-AGI-292.1%mediumEpoch AI
ARC-AGI-259.6%noneEpoch AI
ARC-AGI-293.3%xhighEpoch AI
NYT Connections (extended)98.1%Best of 91xhigh reasoningLech Mazur benchmarks
ARC-AGI-198.5%#5 of 83, top 7%highEpoch AI
ARC-AGI-196.5%lowEpoch AI
ARC-AGI-197.5%maxEpoch AI
ARC-AGI-197.5%mediumEpoch AI
ARC-AGI-186%noneEpoch AI
ARC-AGI-198.5%xhighEpoch AI
CritPt28.9%highEpoch AI
CritPt26.3%lowEpoch AI
CritPt31.7%#4 of 134, top 3%maxEpoch AI
CritPt29.1%mediumEpoch AI
CritPt19.7%noneEpoch AI
CritPt31.4%xhighEpoch AI
Chess Puzzles72%Best of 129maxEpoch AI2026-08-30
EBR-Bench76.2%Best of 24maxEpoch AI2026-08-30
LMArena Hard Prompts1462#44 of 297, top 15%LMArena2026-10-08
Mystery Game Puzzles84%Best of 74maxEpoch AI2026-08-30
DTBench97.1%Epoch AI
DTBench96.5%highEpoch AI
DTBench94.9%lowEpoch AI
DTBench97.1%mediumEpoch AI
DTBench97.3%#5 of 151, top 4%xhighEpoch AI
LMCA64.1%Epoch AI
LMCA63.8%highEpoch AI
LMCA62.2%lowEpoch AI
LMCA62.9%mediumEpoch AI
LMCA64.4%#4 of 125, top 4%xhighEpoch AI
Bench to the Future 30.14#8 of 10, top 80%lowEpoch AI
Epoch Capabilities Index166.45#2 of 213, top 1%Epoch AI2026-09-03

Math

BenchmarkScorePositionSettingSourceDate
FrontierMath (Tiers 1-3)93.7%#2 of 81, top 3%maxEpoch AI2026-08-30
FrontierMath Tier 497.6%#2 of 63, top 4%highEpoch AI2026-08-30
FrontierMath Tier 487.8%lowEpoch AI2026-08-30
FrontierMath Tier 497.6%maxEpoch AI2026-08-30
FrontierMath Tier 497.6%mediumEpoch AI2026-08-30
FrontierMath Tier 482.9%noneEpoch AI2026-08-30
FrontierMath Tier 497.6%xhighEpoch AI2026-08-30
OTIS Mock AIME 2024-2025100%#9 of 173, top 6%maxEpoch AI2026-08-30
ProofBench99%#7 of 77, top 10%Epoch AI
LMArena Math1465#46 of 285, top 17%LMArena2026-10-08
FrontierMath Erdős2.9%Best of 7maxEpoch AI2026-08-28

Knowledge

BenchmarkScorePositionSettingSourceDate
GPQA Diamond95.8%Best of 186maxEpoch AI2026-08-30
Humanity's Last Exam54.8%Best of 41Epoch AI
SimpleQA Verified75.6%Best of 77maxEpoch AI2026-08-30
Vectara Hallucination Rate (lower is better)8.7%#41 of 96, top 43%Vectara Hallucination Leaderboard
LMArena Expert1483#38 of 273, top 14%LMArena2026-10-08

Multimodal

BenchmarkScorePositionSettingSourceDate
LMArena Vision1281#29 of 122, top 24%LMArena2026-10-09
Blueprint-Bench 249.7%#3 of 31, top 10%Epoch AI
Furniture Assembly80%#3 of 31, top 10%maxEpoch AI2026-09-10
LMArena Document1468#13 of 38, top 35%LMArena2026-09-13

Multilingual

BenchmarkScorePositionSettingSourceDate
LMArena Non-English1430#61 of 297, top 21%LMArena2026-10-08
LMArena Chinese1484#61 of 285, top 22%LMArena2026-10-08
LMArena French1456#56 of 223, top 26%LMArena2026-10-08
LMArena German1440#56 of 231, top 25%LMArena2026-10-08
LMArena Japanese1379#79 of 211, top 38%LMArena2026-10-08
LMArena Korean1426#32 of 213, top 16%LMArena2026-10-08
LMArena Russian1436#57 of 283, top 21%LMArena2026-10-08
LMArena Spanish1407#103 of 226, top 46%LMArena2026-10-08

Instruction Following

BenchmarkScorePositionSettingSourceDate
LMArena Instruction Following1450#42 of 298, top 15%LMArena2026-10-08

Long Context

BenchmarkScorePositionSettingSourceDate
LMArena Longer Query1456#44 of 291, top 16%LMArena2026-10-08

Writing & Preference

BenchmarkScorePositionSettingSourceDate
LMArena Text1441#61 of 297, top 21%LMArena2026-10-08
LMArena Creative Writing1418#55 of 295, top 19%LMArena2026-10-08
EQ-Bench Creative Writing2173Best of 115EQ-Bench
LMArena Multi-Turn1448#58 of 295, top 20%LMArena2026-10-08

API pricing by provider

RouteInput $/MOutput $/MCached input $/MChecked
azure$10$50$12026-10-10
bedrock$10$50$12026-10-10
openai$10$50$12026-10-10
openrouter$10$50$12026-10-10

All OpenAI API prices → Estimate your cost →

Compare GPT-6 Astra

Other OpenAI models

Frequently asked questions

How good is GPT-6 Astra?

GPT-6 Astra by OpenAI ranks 1st of 354 ranked models on the Noometry Index as of October 2026, with a score of 70.8. Its strongest category is knowledge, where it ranks 1st. API pricing starts at $10 per million input tokens and $50 per million output tokens, with a 1.05M-token context window.

How much does GPT-6 Astra cost?

GPT-6 Astra costs $10 per million input tokens and $50 per million output tokens on OpenAI's own API, with cached input at $1.

What is GPT-6 Astra's context window?

GPT-6 Astra accepts up to 1.05M tokens of input and can write up to 128K tokens in one response.

Is GPT-6 Astra open source?

No. GPT-6 Astra is proprietary and available only through OpenAI's API and partner platforms.

What are GPT-6 Astra's strengths and weaknesses?

Relative to other ranked models, GPT-6 Astra places best in reasoning, knowledge, coding and lowest in long context, multilingual, instruction following.

What is GPT-6 Astra best at?

Its best category is knowledge, where it ranks 1st on Noometry.