GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.
GPT-6 Astra by openai is offered by 27 providers on this page. Its official list price is $10.00 per 1M input tokens and $50.00 per 1M output tokens. The lowest paid channel is Ofox at $8.00 / $40.00 per 1M, about 1.4× below the list price. That channel is a third-party gateway, so confirm its availability and rate limits before depending on it.
Artificial Analysis rates it 52.7 on the Intelligence Index, with 76.9 for coding and 51 for agentic tasks. Against its lowest blended price of $16.00 per 1M, that is roughly 3 index points per dollar, which is the value ratio the leaderboards rank on. Median output speed is 45 tokens per second, with 376.98s to the first token. Running one task of the Artificial Analysis suite costs about $3.2575, which reflects how many tokens its reasoning consumes rather than the unit price alone.
The context window is 1,050,000 tokens, with an output limit of 128,000 tokens. It supports reasoning, tool use, and structured output. Accepted input modalities are Text, Image, and PDF. Its training knowledge cuts off at 2026-04-30.
Quality is independently evaluated by Artificial Analysis. Speed/latency are model-level medians.
| Provider | Tier | Input | Output | Cache read | Cache write | Context | Output limit | Status |
|---|---|---|---|---|---|---|---|---|
| Ofox | Gateway | $8.00 | $40.00 | $0.80 | $10.00 | 1,050,000 | 128,000 | |
| NanoGPT | Gateway | $10.00 | $50.00 | $1.00 | $12.50 | 1,050,000 | 128,000 | |
| Kilo Gateway | Gateway | $10.00 | $50.00 | $1.00 | $12.50 | 1,050,000 | 128,000 | |
| Venice AI openai-gpt-6-astra | Gateway | $10.00 tiered >272K: $20.00 | $50.00 | $1.00 | $12.50 | 1,050,000 | 128,000 | |
| 302.AI | Gateway | $10.00 | $50.00 | — | — | 1,050,000 | 128,000 | |
| OpenRouter | Gateway | $10.00 tiered >272K: $20.00 | $50.00 | $1.00 | $12.50 | 1,050,000 | 128,000 | |
| Azure | Cloud | $10.00 tiered >272K: $20.00 | $50.00 | $1.00 | $12.50 | 1,050,000 | 128,000 | |
| Azure Cognitive Services | Cloud | $10.00 tiered >272K: $20.00 | $50.00 | $1.00 | $12.50 | 1,050,000 | 128,000 | |
| ZenMux | Gateway | $10.00 tiered >272K: $20.00 | $50.00 | $1.00 | $12.50 | 1,050,000 | 128,000 | |
| Cloudflare AI Gateway | Cloud | $10.00 tiered >272K: $20.00 | $50.00 | $1.00 | $12.00 | 1,050,000 | 128,000 | |
| OpenAIOfficial | First-party | $10.00 tiered >272K: $20.00 | $50.00 | $1.00 | $12.50 | 1,050,000 | 128,000 | |
Sorted by blended price (input×0.75 + output×0.25) asc. The official channel always shows regardless of rank. Whether a gateway's low price is actually usable can't be verified.
Interleaved thinking (reasoning between tool calls) is declared by 1 of 27 providers.
| Mode | Provider | Input | Output | Cache read | Cache write |
|---|---|---|---|---|---|
| fast | OpenAI | $20.00 | $100 | $2.00 | $25.00 |
| pro | OpenAI | — | — | — | — |
| ultrafast | OpenAI | $60.00 | $300 | $6.00 | $75.00 |
| ultrafast | Amazon Bedrock | $66.00 | $330 | $6.60 | $82.50 |
| Name | Conditions | Score | Metric | Source |
|---|---|---|---|---|
| Agents' Last Exam | — | 59.3 | — | Source ↗ |
| OSWorld | dataset: V2-Offline (v2026.08.08) · v2.0 | 72.6 | partial score | Source ↗ |
| ScreenSpot-Pro | variant: no tools | 92.7 | accuracy | Source ↗ |
| AutomationBench | — | 41.4 | success rate | Source ↗ |
| BenchCAD | variant: with tools | 95.9 | geometric overlap | Source ↗ |
| BrowseComp | — | 91.5 | accuracy | Source ↗ |
| Terminal-Bench | v4.0 | 57.9 | success rate | Source ↗ |
| DeepSWE | v1.1 | 74.1 | resolve rate | Source ↗ |
| FrontierCode | dataset: Extended · v1.1 | 64.5 | score | Source ↗ |
| Terminal-Bench Science | v0.1 | 64.6 | success rate | Source ↗ |
| FrontierMath | dataset: Tier 4 · vv2 | 97.6 | accuracy | Source ↗ |
| GPQA Diamond | — | 96 | accuracy | Source ↗ |
| Humanity's Last Exam | variant: with tools | 57.2 | accuracy | Source ↗ |
| ExploitBench | variant: without production safeguards | 100 | success rate | Source ↗ |
| SRE-Bench | variant: single attempt | 88 | success rate | Source ↗ |
| ARC-AGI-3 | harness: Responses API | 99.9 | RHAE | Source ↗ |
| Artificial Analysis Intelligence Index | variant: max · v4.1.1 | 61 | index score | Source ↗ |
| Artificial Analysis Coding Agent Index | harness: Codex · variant: max · v1.4 | 67 | index score | Source ↗ |
The same benchmark scores very differently across harness / dataset, so the qualifying conditions must be shown together.
Individual evaluations run by Artificial Analysis, on the same reasoning tier as the intelligence score above. Each benchmark has its own task set and harness, so rows are not comparable with one another. The Intelligence Index above draws on a different, newer set of evaluations.