Gemini 4 Carbon vs GPT-5.6 Luna
Gemini 4 Carbon leads GPT-5.6 Luna by 22.2 points on the Shogo Agent Index. Gemini 4 Carbon wins 14 and GPT-5.6 Luna wins 0 of 14 shared benchmarks. GPT-5.6 Luna is cheaper for agent workloads.
Add more models to this comparison →Pillars
| Pillar | Gemini 4 Carbon | GPT-5.6 Luna |
|---|---|---|
| Agentic coding | 82 | 59 |
| Long-horizon autonomy | 80 | 61 |
| Knowledge-work agents | 63 | 39 |
| Computer use | 68 | 40 |
| Security & context learning | 63 | 51 |
Shared benchmarks
| Benchmark | Gemini 4 Carbon | GPT-5.6 Luna |
|---|---|---|
| APEX-Agentsindex | 84.2% | 43.0% |
| Vending-Bench 2index | $16,484 | $4,095 |
| CursorBenchindex | 60.4% | 35.9% |
| ARC-AGI | 98.5% | 88.0% |
| ARC-AGI-2 | 94.2% | 59.5% |
| Chess Puzzles | 78.0% | 43.0% |
| CritPt | 36.0% | 20.6% |
| FrontierMath Tier 4 | 95.1% | 61.0% |
| FrontierMath Tiers 1-3 | 90.5% | 82.1% |
| GPQA Diamond | 96.5% | 91.6% |
| OTIS Mock AIME | 100.0% | 98.3% |
| SciCode | 64.2% | 53.6% |
| SimpleBench | 92.6% | 46.8% |
| SimpleQA Verified | 77.7% | 41.0% |
Specs and price
| Gemini 4 Carbon | GPT-5.6 Luna | |
|---|---|---|
| Agent Index | 73.6 (#2) | 51.4 (#34) |
| Agent $/M | $0.84 | $0.10 |
| Input / output $/M | $2.00 / $10.00 | $0.20 / $1.20 |
| Context | 1M | 1.1M |
| Released | Nov 10, 2026 | Jul 9, 2026 |
| Weights | Closed | Closed |
| API hosts | 5 | 5 |
Don’t pick one. Route.
Shogo can run Gemini 4 Carbon and GPT-5.6 Luna side by side and send each step of an agent to the one that’s cheaper for it.
Try both in Shogo