GPT-5.6 Luna vs Grok 4.5: Specs & Benchmark Comparison
GPT-5.6 Luna is developed by OpenAI, while Grok 4.5 comes from xAI. Both were released in July 2026.
The two models share 8 published benchmarks. They split them evenly, 4 to 4. The widest gaps are on DeepSWE, where GPT-5.6 Luna scores 67.2% against 53.0%; DeepSWE 1.1, where GPT-5.6 Luna scores 67.0% against 54.0%. Averaged across everything we track, GPT-5.6 Luna sits at 51.5% and Grok 4.5 at 52.9%.
GPT-5.6 Luna is the cheaper API at $0.20 per million input tokens and $1.20 per million output tokens, roughly 10 times cheaper than Grok 4.5 at $2 and $6. GPT-5.6 Luna takes the larger context window at 1M tokens, compared with 500K for Grok 4.5. Both accept text and images as input.
| Characteristic | GPT-5.6 Luna | Grok 4.5 |
|---|---|---|
| Company | OpenAI | xAI |
| Release Date | July 9, 2026 | July 16, 2026 |
| Parameters | — | — |
| Multimodal | Yes | Yes |
| Context (input) | 1.1M | 500K |
| Context (output) | 128K | — |
| Input Price / 1M | $0.20 | $2.00 |
| Output Price / 1M | $1.20 | $6.00 |
| Average Score | 51.5% | 52.9% |
| Benchmarks | ||
| DeepSWE | 67.2% | 53.0% |
| DeepSWE 1.1 | 67.0% | 54.0% |
| Terminal-Bench 4.0 | 17.3% | 12.4% |
| Artificial Analysis | 51.0% | 54.0% |
| FrontierCode 1.1 | 39.8% | 42.4% |
| SWE-Bench Pro | 62.7% | 65.0% |
| Terminal-Bench 2.1 | 84.7% | 83.0% |
| GPQA | 92.3% | 93.0% |
Visual Benchmark Comparison
Verdict
GPT-5.6 Luna leads in 2 out of 4 comparison categories.
Both models show comparable average scores: GPT-5.6 Luna — 0.5, Grok 4.5 — 0.5.
GPT-5.6 Luna is 5.7x cheaper: input $0.20/1M vs $2.00/1M tokens.
GPT-5.6 Luna supports a larger context: 1M vs 500K tokens.
Both models were released around the same time: 7/9/2026 and 7/16/2026.
More About These Models
Related Comparisons
Frequently Asked Questions
Which is better for coding — GPT-5.6 Luna or Grok 4.5?
Which model is cheaper — GPT-5.6 Luna or Grok 4.5?
Which has a larger context window — GPT-5.6 Luna or Grok 4.5?
The GPT-5.6 Luna and Grok 4.5 comparison is updated for 2026. Data includes benchmark results, API pricing, context window size and other specifications. For more detailed information, visit the GPT-5.6 Luna or Grok 4.5 page. See also the complete list of AI model comparisons.