GPT-5.4 vs GPT-6 Astra
Head to head on 13 shared benchmarks, with price, context window, and release dates.
Benchmark scores
62evals · sort by any column, ⤢ to expand fullscreen
| Model | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| GPT-6 Astra OpenAI | 41.4% | 92.0% | 2164elo | 82.8 | 97.6% | 96.1% | 54.7% | 48.5% | 87.9% | 99.0% | 46.4% | 56.5% | 89.6% | 59.3% | 95.0% | 98.5% | 99.9% | 95.9% | 91.5% | - | - | - | - | 74.1% | - | 100.0% | 42.4% | 53.5% | 53.3% | 37.8% | - | 63.4% | - | - | - | 60.3% | - | - | - | - | - | - | - | - | - | - | 96.3% | - | 0.84 | 72.6% | 92.7% | 85.4% | - | 88.0% | - | - | 57.9% | - | 64.6% | - | 66.6% | - | 74.5% |
| GPT-5.4 OpenAI | 13.1% | 79.0% | 1836elo | 78.3 | 47.6% | 92.0% | 43.7% | 41.3% | 77.5% | 56.0% | 43.3% | 56.6% | 67.4% | - | - | - | - | - | - | 1.47 | 65.3% | 34.1% | 9.0% | - | 1272elo | - | - | - | - | - | 87.2% | - | 73.9% | 78.5% | 92063 | - | 82.0% | 77.5% | 79.3% | 70.2% | 82.6% | 94.1% | 88.1% | 82.8% | 68.3% | 86 | - | 52.2% | - | - | - | - | -0.61 | - | 40.8 | 74.0% | - | 57.6% | - | 80 | - | 87.1% | 65.4% |
2 / 2 models
| Attribute | GPT-5.4 OpenAI | GPT-6 Astra OpenAI |
|---|---|---|
| Description | GPT-5.4 is an AI model from OpenAI. | OpenAI's GPT-6 flagship, launched as its best computer-use model: it drives browsers, spreadsheets and desktop applications rather than answering questions about them. Built on OpenAI's largest training run to date and the first model it designates Critical for cybersecurity capability. Rolling out from Daybreak enterprise access to ChatGPT and the API; Pro, Business and Enterprise plans also get a GPT-6 Astra Pro tier. |
| Family | GPT | GPT |
| Params | - | - |
| Open weights | closed | closed |
| License | proprietary | - |
| Released | 5 Mar 2026 | 3 Sep 2026 |
| Context | 1,050,000 | 1,050,000 |
| Input $/1M | $2.50 | $10.00 |
| Output $/1M | $15.00 | $50.00 |
| Cache read $/1M | $0.250 | $1.000 |
| Cache write $/1M | - | $12.500 |
| Speed | 0 tok/s | 63 tok/s |
| Latency | 0.00 s | 198.54 s |
| Modalities | text, image, file | file, image, text |
| Providers | - | openai |
| Publisher | OpenAI | OpenAI |
| Reported scores | GPQA Diamond92.0% (AA) SciCode56.6% (AA) τ²-bench (Tau²-bench)87.1% (AA) Terminal-Bench (Hard)57.6% (AA) Vibe Code Bench v1.167.4% |
FAQ
- Is GPT-5.4 or GPT-6 Astra better?
- Across the 13 benchmarks both models report on Sophon, GPT-6 Astra leads on 12 of them. Which one is "better" depends on the benchmark - the table above shows every shared score.
- How do GPT-5.4 and GPT-6 Astra compare on GPQA Diamond?
- GPT-5.4 scores 92.0% and GPT-6 Astra scores 96.1% on GPQA Diamond.
- Which is cheaper, GPT-5.4 or GPT-6 Astra?
- GPT-5.4 is cheaper at $15.00 per million output tokens against $50.00 for GPT-6 Astra - about 3.3x.
- When were GPT-5.4 and GPT-6 Astra released?
- GPT-5.4 was released 5 Mar 2026 by OpenAI. GPT-6 Astra was released 3 Sep 2026 by OpenAI.
Other head-to-heads
56Gemini 3.1 Pro Preview vs GPT-5.434 shared evalsClaude Opus 4.6 vs GPT-5.433 shared evalsGPT-5.2 vs GPT-5.429 shared evalsGPT-5.4 vs GPT-5.528 shared evalsClaude Opus 4.8 vs GPT-5.428 shared evalsClaude Opus 4.5 vs GPT-5.428 shared evalsGPT-5.4 vs Kimi K2.528 shared evalsClaude Sonnet 4.6 vs GPT-5.428 shared evalsClaude Opus 4.7 vs GPT-5.428 shared evalsGPT-5 Mini vs GPT-5.427 shared evalsGPT-5.4 Mini vs GPT-5.427 shared evalsClaude Sonnet 4.5 vs GPT-5.426 shared evalsClaude Fable 5 vs GPT-5.426 shared evalsGLM 5.1 vs GPT-5.426 shared evalsGPT-5.4 vs Grok 4.1 Fast25 shared evalsGPT-5.4 vs Qwen3.6 Plus24 shared evalsDeepSeek V3.2 vs GPT-5.424 shared evalsGPT-5 Nano vs GPT-5.422 shared evalsClaude 4 Sonnet vs GPT-5.422 shared evalsGLM 4.7 vs GPT-5.421 shared evalsGemini 2.5 Pro vs GPT-5.420 shared evalsGPT-5 vs GPT-5.419 shared evalsGPT-5.5 vs GPT-6 Astra19 shared evalsClaude Opus 4.8 vs GPT-6 Astra19 shared evalsClaude Fable 5 vs GPT-6 Astra18 shared evalsGPT-5.4 vs gpt-oss-120b18 shared evalsGemini 3.1 Pro Preview vs GPT-6 Astra16 shared evalsGPT-4o vs GPT-5.414 shared evalsClaude Opus 4.7 vs GPT-6 Astra14 shared evalsGLM 5.1 vs GPT-6 Astra14 shared evalsGPT-5.2 vs GPT-6 Astra13 shared evalsClaude Opus 4.6 vs GPT-6 Astra13 shared evalsGPT-4.1 vs GPT-5.413 shared evalsGPT-6 Astra vs Kimi K2.513 shared evalsClaude Sonnet 4.6 vs GPT-6 Astra13 shared evalsGPT-4.1 Mini vs GPT-5.412 shared evalsGPT-5 Mini vs GPT-6 Astra12 shared evalsGPT-5 vs GPT-6 Astra12 shared evalsGPT-4o-mini vs GPT-5.412 shared evalsGPT-4.1 Nano vs GPT-5.412 shared evalsGPT-5.4 Mini vs GPT-6 Astra12 shared evalsClaude Sonnet 4.5 vs GPT-6 Astra11 shared evalsClaude Opus 4.5 vs GPT-6 Astra11 shared evalsGPT-6 Astra vs Qwen3.6 Plus11 shared evalsGPT-5 Nano vs GPT-6 Astra10 shared evalsGemini 2.5 Pro vs GPT-6 Astra10 shared evalsGPT-6 Astra vs Grok 4.1 Fast10 shared evalsDeepSeek V3.2 vs GPT-6 Astra9 shared evalsGLM 4.7 vs GPT-6 Astra9 shared evalsClaude 4 Sonnet vs GPT-6 Astra8 shared evalsGPT-4.1 Mini vs GPT-6 Astra6 shared evalsGPT-4.1 Nano vs GPT-6 Astra6 shared evalsGPT-6 Astra vs gpt-oss-120b6 shared evalsGPT-4.1 vs GPT-6 Astra5 shared evalsGPT-4o-mini vs GPT-6 Astra4 shared evalsGPT-4o vs GPT-6 Astra4 shared evals
Comparing something else? Build your own side-by-side.