GPT-6.2 Astra
A frontier reasoning model for coding, computer use, and professional work such as slides, spreadsheets, and documents.
Shogo Agent Index
UnrankedNot ranked yet: 0 agentic benchmarks measured, and we need at least 2 plus an overall capability score. It will be ranked as results are published.
Agentic coding
88
estimated
Long-horizon autonomy
84
estimated
Knowledge-work agents
67
estimated
Computer use
79
estimated
Security & context learning
66
estimated
Price
- Input
- $8.00/M
- Output
- $40.00/M
- Cache read
- $0.40/M
- 3:1 blend
- $16.00/M
- Agent-effective
- $3.37/M
Cheapest host for agents: Amazon Bedrock at $3.37/M.
Agent-effective = 20 input tokens per output token, 85% cache hits. Why.
Agentic benchmarks
| Benchmark | Score | Normalised | Index weight | Harness / source |
|---|---|---|---|---|
| SWE-bench Verified | est. | 8% | predicted 92 | |
| Terminal-Bench 2.0 | est. | 8% | predicted 98 | |
| DeepSWE | est. | 6% | predicted 89 | |
| FrontierSWE | est. | 5% | predicted 79 | |
| CursorBench | est. | 4% | predicted 63 | |
| GSO-Bench | est. | 4% | predicted 89 | |
| METR Time Horizons | est. | 10% | predicted 90 | |
| Vending-Bench 2 | est. | 6% | predicted 97 | |
| PostTrainBench | est. | 4% | predicted 46 | |
| APEX-Agents | est. | 8% | predicted 84 | |
| GDPval | est. | 7% | predicted 79 | |
| Remote Labor Index | est. | 5% | predicted 36 | |
| DeepResearch Bench | est. | 5% | predicted 55 | |
| OSWorld 2.0 | est. | 6% | predicted 66 | |
| OSWorld | est. | 4% | predicted 98 | |
| Cybench | est. | 5% | predicted 97 | |
| CL-bench | est. | 5% | predicted 35 |
Other benchmarks
| Benchmark | Category | Score | Source |
|---|---|---|---|
| Chess Puzzles | Reasoning | 78.0% | source |
| FrontierMath Tiers 1-3 | Math | 92.6% | source |
| FrontierMath Tier 4 | Math | 100.0% | source |
| GPQA Diamond | Reasoning | 96.8% | source |
| MirrorCode | Coding | 67.8% | source |
| OTIS Mock AIME | Math | 100.0% | source |
| SimpleQA Verified | Knowledge | 78.5% | source |
API hosts
Live prices and uptime for every host serving GPT-6.2 Astra via OpenRouter.
| Host | Input | Output | Cache read | Agent $/M | Uptime 24h | Quant |
|---|---|---|---|---|---|---|
| Amazon Bedrock | $8.00 | $40.00 | $0.40 | $3.37 | 99.9% | — |
| Azure | $8.00 | $40.00 | $0.40 | $3.37 | 100.0% | — |
| OpenAI | $8.00 | $40.00 | $0.40 | $3.37 | 99.9% | — |
| OpenAI | $16.00 | $80.00 | $0.80 | $6.74 | 100.0% | — |
| OpenAI | $24.00 | $120.00 | $1.20 | $10.11 | 100.0% | — |