NEWHow AI models draw the world map · Gecko Tests
Beta
Home/Models/GPT-6 Astra
OpenAI logo

GPT-6 Astra

by OpenAI · Released Sep 2026

Multimodal1M Context
96.3
avg score
Rank #4
Compare
Top 1% of all models
Context
1.1M tokens (~525 books)
Input $/1M
$10.00
Output $/1M
$50.00
Type
multimodal
License
Proprietary
Benchmarks
31 tested
Data as of
About

GPT-6 Astra is OpenAI's flagship model for demanding end-to-end work. It is suited for advanced analysis, software engineering, deep research, scientific work, and document creation, with particular strengths in long-horizon...

Tested on 31 benchmarks · BenchGecko score 96.3. Top scores: OTIS Mock AIME 2024-2025 (100.0%), Proofbench (99.0%), ARC-AGI (98.5%).

Looking for similar performance at lower cost?
Step 3.5 Flash scores 89.5 (93% as good) at $0.10/1M input · 99% cheaper
Capabilities
coding
66.6
#23 globally
reasoning
96.8
#1 globally
math
97.1
#2 globally
knowledge
72.3
#14 globally
agentic
64.7
#9 globally
general
74.4
#4 globally
speed
63.9
#2 globally
Benchmark Scores
Compare All
Tested on 31 benchmarks · Ranked across 7 categories
Score Distribution (all 312 models)
0255075100
▲ You are here
codingCompare coding →
WeirdML

Unusual and adversarial machine learning challenges. Tests robustness of reasoning about edge cases in ML systems.

93.6·
Deepswe
74.1·
Frontierswe
65.5·
reasoningCompare reasoning →
ARC-AGI

Abstraction and Reasoning Corpus. Tests fluid intelligence through novel visual pattern recognition puzzles. Core measure of general intelligence.

98.5·
ARC-AGI-2

ARC-AGI 2, harder sequel to ARC. More complex abstract reasoning patterns that test generalization ability beyond training data.

95.0·
mathCompare math →
OTIS Mock AIME 2024-2025

Mock AIME (American Invitational Mathematics Exam) problems from OTIS. Tests mathematical competition performance.

100.0·
FrontierMath-Tier-4-v2-Private
97.6·
FrontierMath-Tiers-1-3-v2-Private
93.7·
Quick compare:Anthropic logovs Claude Opus 5.5DeepSeek logovs DeepSeek V3.2 SpecialeAnthropic logovs Claude Mythos Preview
Excellent (85+) Good (70-85) Average (50-70) Below (<50)
Model Family · OpenAI GPT-6
Sep 2026
96.3
$10.00/M in1.1Mctx31 benchmarks
Sep 2026
$5.00/M in(-5)1.1Mctx
Sep 2026
$10.00/M in(+5)1.1Mctx
Sep 2026
$5.00/M in(-5)1.1Mctx
Sep 2026
$0.10/M in(-4.90)1.1Mctx7 benchmarks
Sep 2026
$0.05/M in(-0.05)1.1Mctx
Sep 2026
$0.10/M in(+0.05)1.1Mctx
Sep 2026
$0.05/M in(-0.05)1.1Mctx
Sep 2026
$2.00/M in(+1.95)1.1Mctx
Sep 2026
$1.00/M in(-1)1.1Mctx
Sep 2026
$2.00/M in(+1)1.1Mctx
Sep 2026
$1.00/M in(-1)1.1Mctx
See the full GPT-6 family →
Recently Happened
GPT-6 Astra added
Sep 5, 2026
Similar Models
Anthropic logo
Claude Opus 5.5
Anthropic
96.8$4.00/1M
DeepSeek logo
DeepSeek V3.2 Speciale
DeepSeek
95.2$0.40/1M
Anthropic logo
Claude Mythos Preview
Anthropic
99.8n/a
Links
Info
OpenAIPricing explorerDevelopers · API
Research
Technical Report
Documentation
API DocsPlayground
Community
@OpenAI
BenchGecko API
gpt-6-astra
Specifications
  • Typemultimodal
  • Context1.1M tokens (~525 books)
  • ReleasedSep 2026
  • LicenseProprietary
  • StatusActive
  • Cost / Message~$0.070
Available On
OpenAI logoOpenAI$10.00
Categories
codingreasoningmathknowledgeagenticgeneralspeed
Learn More
context-windowtransformertokens
Share & Export
Tweet
Related Models
Anthropic logoClaude Opus 5.5
96.8
DeepSeek logoDeepSeek V3.2 Speciale
95.2
Anthropic logoClaude Mythos Preview
99.8
OpenAI logoGPT-5.5 Pro
99.9
GPT-6 Astra is a proprietary multimodal AI model by OpenAI, released in September 2026. It has an average benchmark score of 96.3. Context window: 1M tokens.

Related Models

Claude Opus 5.5 · AnthropicDeepSeek V3.2 Speciale · DeepSeekClaude Mythos Preview · AnthropicGPT-5.5 Pro · OpenAIStep 3.5 Flash · stepfun

Benchmarks

OTIS Mock AIME 2024-2025ProofbenchARC-AGIFrontierMath-Tier-4-v2-PrivateArtificial Analysis · GPQA Diamond

Related Pages

OpenAI · ProviderOpenAI · EconomyAll ModelsCompare ModelsPricingDevelopers · APIContext Window · Glossary

Key facts · as of 2026-10-05

  • GPT-6 Astra by OpenAI. BenchGecko score 96.3, rank 4 of 312 scored models (normalized average of public benchmark scores).
  • List price $10.00 input · $50.00 output per 1M tokens (as of 2026-10-05).
  • Sold by 7 providers (as of 2026-10-05): OpenAI $5.00 in / $25.00 out · Azure $10.00 in / $50.00 out · OpenAI $10.00 in / $50.00 out · Amazon Bedrock $11.00 in / $55.00 out · Azure $11.00 in / $55.00 out · and 2 more. Every provider

How to cite · data as of 2026-10-05

GPT-6 Astra · benchmarks, pricing and providers. BenchGecko, data as of 2026-10-05. https://benchgecko.ai/model/gpt-6-astra

Credit "Source: BenchGecko" with a link. Prices per provider and Gecko Tests are BenchGecko data (CC BY 4.0); benchmark scores keep their original source, listed in the JSON. JSON · llms.txt · MCP