
Automatically append previous messages to maintain multi-turn context. May increase token usage.
How can I help you today?
README
Complete guide to using gpt-6-2-astra
GPT-6.2 Astra API for Complex Engineering, Science and Professional Work

curl https://api.kie.ai/codex/v1/responses \-H "Authorization: Bearer $KIE_API_KEY" \-H "Content-Type: application/json" \-d '{"model": "gpt-6-2-astra","input": "Design an idempotent webhook handler for payment retries.","reasoning": { "effort": "medium" },"stream": false}'
Key Capabilities of GPT-6.2 Astra


Cleaner Code, Web Frontends and 3D Design

1,050,000-Token Context and Up to 128,000 Output Tokens

Five Reasoning Effort Levels From Low to Max

Scientific Research and Defensive Security

Responses API Tools and Preserved Reasoning
GPT-6.2 Astra Specifications
GPT-6.2 Astra specifications from OpenAI's documentation, alongside Kie.ai endpoint details.
| Specification | GPT-6.2 Astra |
|---|---|
| Developer | OpenAI |
| Model ID | gpt-6-2-astra on Kie.ai (gpt-6.2-astra also accepted) |
| Release date | November 17, 2026 |
| Modalities | Text and image input; text output (audio and video unsupported) |
| Context window | 1,050,000 tokens (up to 922,000 input tokens) |
| Max output tokens | 128,000 tokens |
| Reasoning effort | low, medium, high, xhigh, max |
| Kie.ai endpoint & tools | POST /codex/v1/responses; custom function calling or web_search (one per request); text, input_image and input_file |
| Kie.ai features | Streaming (default), function calling or web search (one per request), image and file input, prompt caching |
| Unsupported parameters and features | Parameters: temperature, top_p, logprobs, top_logprobs; features: fine-tuning, predicted outputs |
| Knowledge cutoff | May 18, 2026 |
GPT-6.2 Astra Benchmark Highlights
Evaluation results reported by OpenAI in its November 17, 2026 release announcement, launch posts and System Card.
| Benchmark | GPT-6.2 Astra Result | What It Measures |
|---|---|---|
| Agents' Last Exam | +5.2 points over GPT-6 Astra (new high) | Long-horizon multistep professional workflows |
| Terminal-Bench-Science 0.1 | +9 percentage points over GPT-6 Astra (new high) | Scientific work carried out in a terminal |
| GDPval v2 | +6.2 percentage points over GPT-6 Astra with 22% fewer output tokens | Multistep consulting, finance and design assignments |
| AA-AnalystAgent | 62.5% of questions answered correctly on all 5 independent attempts (new high) | Financial and expenditure analysis over real spreadsheets and reports |
| AA-3D | 1,318 Elo, ranked first (GPT-6.2 Sol: 1,259 Elo) | Blinded human preference on interactive 3D browser builds |
| BenchCAD 2.1 Core | 32% improvement over GPT-6 Astra | Building 3D parts and assemblies from engineering drawings |
| Cybersecurity evaluations | System Card targets: 36.4% solved vs 18.2% for GPT-6 Astra; ExploitBench Internal Port: +22 points over GPT-6 Astra | Separate security evaluations reported by OpenAI |
How to Work With GPT-6.2 Astra
- 1
Step 1: Choose Between GPT-6.2 Astra and GPT-6.2 Sol
Use GPT-6.2 Astra for your hardest projects across engineering, scientific research, professional deliverables, frontends and 3D design. For day-to-day coding, OpenAI recommends starting with GPT-6.2 Sol, its efficient, versatile model for everyday work. - 2
Step 2: Set Reasoning Effort From Low to Max
On Kie.ai, pass reasoning.effort as low, medium, high, xhigh or max in /codex/v1/responses. OpenAI recommends starting with medium for a new integration, or switching legacy none and minimal settings to low. Omit temperature, top_p, logprobs and top_logprobs. - 3
Step 3: Keep Prompt Prefixes Unchanged
Starting with GPT-6.2, OpenAI's API drops reasoning items when it detects mid-conversation prefix edits (such as changes to instructions or tool calls), a model downgrade, or a change in the end user or organization. OpenAI recommends keeping prompt prefixes unchanged and appending new instructions to help preserve reasoning across turns and keep earlier input eligible for prompt caching. - 4
Step 4: Call /codex/v1/responses for Tools and Multimodal Inputs
On Kie.ai, send requests to POST /codex/v1/responses with custom function definitions or {"type":"web_search"} (choose one tool type per request). Because GPT-6.2 Astra writes more concisely and directly by default, review legacy system prompts that repeat "be concise" instructions.