GPT-5.6 LunavsGrok 4.5
GPT-5.6 Luna | Grok 4.5 | |
|---|---|---|
| Benchmarks | ||
Nonsense detection BullshitBench v2Given a confidently-worded but nonsensical prompt, does the AI spot that it makes no sense and push back — instead of playing along and inventing an answer? The score is how often it clearly called out the nonsense. Higher is better. | 40% | 54%Best |
Prompt injection robustness Gray Swan IPI · k = 1Attackers hide malicious instructions inside content the AI reads — a web page, an email, a document — and try to hijack what it does. Gray Swan's indirect prompt injection benchmark measures how often such an attack succeeds when the attacker gets a single try. Lower is better. | 8.3% | 13.4%Best |
Prompt injection robustness Gray Swan IPI · k = 10Attackers hide malicious instructions inside content the AI reads — a web page, an email, a document — and try to hijack what it does. This variant gives the attacker 10 tries and counts an attack as successful if any of them works. Lower is better. | 38.6% | 54.2%Best |
Prompt injection robustness Gray Swan IPI · k = 15Attackers hide malicious instructions inside content the AI reads — a web page, an email, a document — and try to hijack what it does. This variant gives the attacker 15 tries and counts an attack as successful if any of them works. Lower is better. | 43.9% | 60.8%Best |
Agentic coding SWE-Bench ProCan the AI fix real bugs in real software? It's handed actual problems from open-source projects and has to write code that genuinely solves them. Higher is better. | — | 64.7% |
Multilingual coding SWE-Bench MultilingualLike SWE-Bench, but the coding problems span many programming languages, not just one. Tests how broadly the AI can code. Higher is better. | — | 78% |
Agentic coding CursorBench v3.2Cursor's own test of harder, real-world coding tasks inside a code editor, on the refreshed v3.2 task set. Scores aren't comparable with v3.1. Higher is better. | 61.1% | 66.7%Best |
Agentic coding DeepSWE 1.1Artificial Analysis' independent test of deep, agentic software-engineering work — the AI has to plan and carry out substantial coding tasks end to end. (Version 1.1 of the test.) Higher is better. | — | 54% |
Agentic coding DeepSWE 1.0Artificial Analysis' independent test of deep, agentic software-engineering work — the AI has to plan and carry out substantial coding tasks end to end. Higher is better. | — | 62% |
Next.js coding Next.js EvalsVercel's open eval of how well AI coding agents build and migrate real Next.js apps — measured as the share of tasks the agent completes successfully. Higher is better. | — | 83% |
Agentic computer work Frontier-Bench v0.1A hard, ever-evolving set of real computer tasks — coding, system administration, data work, and more — that an AI agent has to complete on its own. Run by the Harbor / Laude Institute team as the successor to Terminal-Bench (v0.1 is the first release of the task set). The score is the share of tasks solved. Higher is better. | 14.3% | 17.8%Best |
Agentic terminal coding Terminal-Bench 2.1Can the AI work in a command-line terminal — running commands and finishing technical setup tasks the way a developer would? Higher is better. | 82.5% | 83.3%Best |
Agentic legal work Harvey's Legal Agent BenchmarkHarvey's test of whether an AI agent can complete real legal work — drafting and reviewing documents, working with spreadsheets and presentations, and navigating files the way a lawyer's assistant would. Higher is better. | — | 12.92% |
Medical admin work MedScribeCan the AI support doctors with their administrative work, like notes and paperwork? Created by Vals AI. Higher is better. | — | 86.88% |
Community preference Arena Elo (Text)Real people chat with two anonymous AIs side by side and vote for the answer they prefer. Votes become a chess-style Elo rating on arena.ai — it measures which AI people actually like, not test scores. Higher is better. | — | 1468 |
Community preference (code) Arena Elo (Code)Like the text arena, but people vote on which AI writes better code. The votes become a chess-style Elo rating on arena.ai. Higher is better. | 1523 | 1549Best |
| Overview | ||
| Company | OpenAI | SpaceXAI |
| Release date | Jun 26 2026 | Jul 8 2026 |
| Access | Proprietary | Proprietary |
Which is better: GPT-5.6 Luna or Grok 4.5?
Grok 4.5 leads GPT-5.6 Luna on 5 of the 8 benchmarks they both report. GPT-5.6 Luna shipped 12 days before Grok 4.5, so benchmark comparisons should account for the intervening progress.
Published specifications for these two models are limited — see each model page for the latest details.
On BullshitBench v2, Grok 4.5 leads at 54% vs GPT-5.6 Luna at 40%. On Gray Swan IPI · k = 1, GPT-5.6 Luna leads at 8.3% vs Grok 4.5 at 13.4%. On Gray Swan IPI · k = 10, GPT-5.6 Luna leads at 38.6% vs Grok 4.5 at 54.2%. On Gray Swan IPI · k = 15, GPT-5.6 Luna leads at 43.9% vs Grok 4.5 at 60.8%. On CursorBench v3.2, Grok 4.5 leads at 66.7% vs GPT-5.6 Luna at 61.1%. On Frontier-Bench v0.1, Grok 4.5 leads at 17.8% vs GPT-5.6 Luna at 14.3%. On Terminal-Bench 2.1, Grok 4.5 leads at 83.3% vs GPT-5.6 Luna at 82.5%. On Arena Elo (Code), Grok 4.5 leads at 1549 vs GPT-5.6 Luna at 1523.
Frequently asked questions
GPT-5.6 Luna was released by OpenAI on Jun 26 2026.
Grok 4.5 was released by SpaceXAI on Jul 8 2026.
Grok 4.5 leads on CursorBench v3.2 — GPT-5.6 Luna 61.1% vs Grok 4.5 66.7%.
Other comparisons
GPT-5.6 LunavsClaude Opus 5Grok 4.5vsClaude Opus 5GPT-5.6 LunavsGemini 3.6 FlashGrok 4.5vsGemini 3.6 FlashGPT-5.6 LunavsMuse GlimmerGrok 4.5vsMuse GlimmerGPT-5.6 LunavsDeepSeek-V4-Flash-0731Grok 4.5vsDeepSeek-V4-Flash-0731GPT-5.6 LunavsMistral Medium 3.5Grok 4.5vsMistral Medium 3.5GPT-5.6 LunavsKimi K3Grok 4.5vsKimi K3