DeepSeek V4.1-Flash vs GPT-6 Astra
DeepSeek V4.1-Flash is from DeepSeek, released September 10, 2026. GPT-6 Astra is from OpenAI, released September 3, 2026.
- LMArena Elo1476 vs 1476. Both rated 1476 on LMArena.
- Input per 1M$0.30 vs $10. DeepSeek V4.1-Flash costs 97% less per million input tokens.
- Output per 1M$1.20 vs $50. DeepSeek V4.1-Flash costs 98% less per million output tokens.
- Context window1M vs 1.05M. GPT-6 Astra takes 5.0% more context in one request.
Prices are the vendors’ published API rates; Elo is LMArena’s text board. Each figure is dated on its model page.
Interactive Comparison
Loading chart...
$0.30 / $1.20 per million tokens | $10 / $50 per million tokens | |
|---|---|---|
| Identity | ||
| Developer | DeepSeek | OpenAI |
| Released | Sep 2026 | Sep 2026 |
| Status | GA | GA |
| Licence | MIT | Proprietary |
| Self-hostable | Yes - open weights on Hugging Face | No |
| Cost | ||
| Blended $/1M tokensinput × 0.75 + output × 0.25 | $0.525 / 1M tokens | $20.000 / 1M tokens |
| Input price | $0.300 / 1M tokens | $10.000 / 1M tokens |
| Output price | $1.200 / 1M tokens | $50.000 / 1M tokens |
| Cached input | $0.006 / 1M tokens (peak); $0.003 off-peak | $1 per 1M tokens |
| Batch discount | Off-peak billing: 50% off all rates outside 01:00-04:00 and 06:00-10:00 UTC, Monday to Friday (weekends and Chinese public holidays are off-peak in full) | 50% (Batch and Flex) |
| Capacity | ||
| Context window | 1M tokens | 1,050,000 tokens |
| Max output | 384k tokens | 128k tokens |
| Long-context surcharge | None (flat rate across the 1M window) | Yes - prompts over 272K input tokens are priced at 2x input and cache rates and 1.5x output for the full request |
| Capability | ||
| Vision in | Yes | Yes |
| Audio in | No | No |
| Function calling | Yes | — |
| Structured output | Yes - JSON output | — |
| Extended reasoning | Yes - thinking (default) and non-thinking modes | — |
| Web search | No | — |
| Code execution | No | — |
| Access | ||
| API | Yes | Yes |
| Chat app | Yes | Yes |
| Cloud marketplaces | Not offered | — |
| Fine-tuning | Not offered (weights are open) | — |
| Measured quality | ||
| LMArena Elo | 1476 (checked Oct 2026) | 1476 (checked Oct 2026) |
| LMArena rank | Rank 35 | Rank 37 |
| GPQA Diamond | 90 · max effort | 96 |
| Elo per dollarLMArena Elo ÷ blended $/1M tokens | 2811 | 74 |
