OpenAI · Released Sep 4, 2026
GPT-6 Astra
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...
- #1Vending-Bench 2
- #1ARC-AGI-2
- #1ARC-AGI-1
- #1Chess Puzzles
- Released
- Sep 4 2026
- Context window
- 1.05M 128K out
- Input / 1M tokens
- $10.00 from $5.00
- Output / 1M tokens
- $50.00 from $25.00
Where to run it
3 providers compared
Cheapest: OpenAI at $5.00 / $25.00 per 1M input / output tokens — 92% less than the priciest option.
| Provider | Input | Output | Cache read | Context | Max output | Uptime | vs. best |
|---|---|---|---|---|---|---|---|
| OpenAI | $5.00 | $25.00 | $0.50 | 1.05M | 128K | 99.9%24h | Cheapest |
| Azure | $10.00 | $50.00 | $1.00 | 1.05M | 128K | 100%24h | +100% |
| OpenAI | $10.00 | $50.00 | $1.00 | 1.05M | 128K | 100%24h | +100% |
| Amazon Bedrockus-west-2 | $11.00 | $55.00 | $1.10 | 1.05M | 128K | 100%24h | +120% |
| Azureus | $11.00 | $55.00 | $1.10 | 1.05M | 128K | 100%24h | +120% |
| OpenAIfast | $20.00 | $100.00 | $2.00 | 1.05M | 128K | 100%24h | +300% |
| OpenAIultrafast | $60.00 | $300.00 | $6.00 | 1.05M | 128K | 100%24h | +1100% |
Value
Price vs. performance
GPT-6 Astra Other OpenAI Other labs Value frontier · GPQA Diamond vs. input price per 1M tokens (log scale) · Drag to pan, pinch or Ctrl + scroll to zoom, double-click to zoom in.
Benchmarks
26 benchmark scores
| Benchmark | Score | Rank | Leader |
|---|---|---|---|
| FrontierSWESoftware engineering · Ultra-long-horizon engineering tasks with a 20-hour budget each | 65.5% | #1of 19 | This model |
| DeepSWESoftware engineering · Original long-horizon tasks across active open-source repos | 74.1% | #1of 26 | This model |
| FrontierCodeSoftware engineering · Mergeable fixes for hard open-source issues | 53.3% | #4of 41 | Claude Opus 5.554.6% |
| MirrorCodeSoftware engineering · Reimplement whole programs end-to-end without the source | 46.7% | #4of 9 | Claude Opus 5.577.4% |
| GSOSoftware engineering · Optimize real code to make it significantly faster | 79.4% | #2of 20 | Claude Fable 5.188.2% |
| SciCodeSoftware engineering · Research-coding problems from the natural sciences | 56.5% | #20of 93 | Claude Opus 5.566.9% |
| WeirdML v3Software engineering · Unfamiliar ML and data tasks with vague goals | 42.2% | #1of 22 | This model |
| APEX-AgentsAgents & computer use · Long professional tasks in banking, consulting and law | 64.7% | #8of 51 | Gemini 4 Argon82.4% |
| Remote Labor IndexAgents & computer use · Real paid freelance projects done end-to-end | 20.8% | #2of 13 | Claude Opus 5.521.3% |
| Vending-Bench 2Agents & computer use · Run a simulated vending business for a year (final balance) | $15,515 | #1of 61 | This model |
| GDP.pdfWriting & content · Answer questions about real professional PDFs | 34.2% | #2of 38 | GPT-6.1 Sol35.0% |
| SimpleQA VerifiedWriting & content · Factual accuracy on 1,000 short questions | 75.6% | #2of 66 | Gemini 4 Argon76.0% |
| WebDev ArenaDesign & visual · People vote on which model builds the better website (Elo) | 1800 | #2of 88 | Claude Opus 5.51820 |
| Blueprint-Bench 2Design & visual · Turn apartment photos into 2D floor plans | 49.7% | #3of 32 | Gemini 4 Argon54.4% |
| Furniture AssemblyDesign & visual · Spot mistakes in photos of IKEA builds | 80.0% | #2of 34 | Claude Opus 5.583.3% |
| GPQA DiamondReasoning & knowledge · PhD-level biology, chemistry and physics questions | 95.8% | #2of 92 | Gemini 4 Argon96.4% |
| Humanity's Last ExamReasoning & knowledge · Expert questions across 100+ subjects | 54.8% | #1of 23 | This model |
| ARC-AGI-2Reasoning & knowledge · Few-shot abstract reasoning on grid puzzles | 95.0% | #1of 63 | This model |
| ARC-AGI-1Reasoning & knowledge · Pattern induction from input–output examples | 98.5% | #1of 63 | Gemini 4 Argon98.5% |
| CritPtReasoning & knowledge · Open-ended, research-level physics problems | 31.7% | #2of 96 | GPT-5.6 Sol32.3% |
| Chess PuzzlesReasoning & knowledge · Find the single best move in 100 new puzzles | 76.0% | #1of 70 | This model |
| LMCAReasoning & knowledge · Judge the quality of conceptual arguments like experts | 64.4% | #5of 96 | Claude Sonnet 5.568.3% |
| FrontierMathMaths · Expert-written problems, undergrad to early research (tiers 1–3) | 93.7% | #1of 72 | GPT-6.1 Sol93.7% |
| FrontierMath Tier 4Maths · Exceptionally hard research-level problems | 97.6% | #2of 62 | GPT-6.1 Sol100.0% |
| ProofBenchMaths · Graduate maths proofs verified in Lean 4 | 99.0% | #4of 70 | Claude Sonnet 5.5100.0% |
| OTIS Mock AIMEMaths · Competition-style problems harder than MATH Level 5 | 100.0% | #1of 90 | Gemini 4 Argon100.0% |
Specs
Details
- Model ID
openai/gpt-6-astra- Developer
- OpenAI
- Release date
- Sep 4, 2026 ·
- Context window
- 1,050,000 tokens
- Max output
- 128,000 tokens
- API price
- $10.00 / $50.00 per 1M input / output tokens
- Providers
- 3
- Input types
- file, image, text
- Reasoning
- Yes
- Tool calling
- Yes
- Open weights
- No
- Free tier
- No
Release date, prices and providers from OpenRouter's catalog; buying directly from a provider can cost a little differently. Uptime is the share of OpenRouter requests each provider served over the last day. Benchmark scores are the best result per benchmark from Epoch AI, licensed CC BY 4.0. Rank counts every model tracked here with a score.