GPT-6 Astra
OpenAI
88.0#5 of 427 overall
- Input / output
- US$ 10,00 / US$ 50,00 US$ per 1M tokens
- Context
- 1.1M tokens
- Release
- 09/04/2026 release date
- Weights
- Closed
- Speed
- 88 tok/s first token in 184.03 s
Scores by source
Scores by sourcepercentile = % of ranked models it beats| Metric | Value | Percentile | Source |
|---|---|---|---|
| LiveBench weight 3 | 85.6 | 97% | LiveBench |
| AA Intelligence Index weight 3 | 56.9 | 95% | Artificial Analysis |
| Arena text (Elo) weight 2 | 1441 | 71% | LMArena |
| Arena WebDev (Elo) weight 1 | 1773 | 99% | LMArena |
| Arena vision (Elo) weight 0.5 | 1280 | 73% | LMArena |
| Humanity's Last Exam weight 1 | 54.7 | 99% | Artificial Analysis |
| Terminal-Bench weight 1 | 59.1 | 97% | Artificial Analysis |
| GPQA Diamond | 96.1 | 100% | Artificial Analysis |
| Long-context reasoning | 80.7 | 89% | Artificial Analysis |
| LiveBench long agentic | 77.5 | 97% | LiveBench |
| LiveBench reasoning | 92.7 | 97% | LiveBench |
| LiveBench coding | 93.7 | 77% | LiveBench |
| LiveBench browser | 91.3 | 94% | LiveBench |
| LiveBench data analysis | 83.0 | 97% | LiveBench |
| LiveBench research | 85.2 | 94% | LiveBench |
| LiveBench design | 82.0 | 74% | LiveBench |
By category
| Category | Score | Rank | Metrics |
|---|---|---|---|
| Agentic | 81.3 | #1 | LiveBench long agentic, Terminal-Bench |
| Coding | 75.5 | #6 | LiveBench coding, Arena WebDev (Elo) |
| Reasoning | 86.3 | #1 | LiveBench reasoning, Humanity's Last Exam, GPQA Diamond |
| Data analysis | 73.4 | #2 | LiveBench data analysis |
| Research | 77.6 | #3 | LiveBench research, Long-context reasoning |
| Design | 74.4 | #6 | LiveBench design, Arena WebDev (Elo) |
| Browser use | 71.8 | #3 | LiveBench browser |
| Chat | 60.6 | #57 | Arena text (Elo) |
| Vision | 61.6 | #29 | Arena vision (Elo) |
History
RankingIA score on each collection day (hover to see the rank).
Daily collection started on 10/30/2026. The chart appears from the second day the model is ranked.