Compare · ModelsLive · 2 picked · head to head
Gemini 2.5 Pro vs GPT-5.6 Luna
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
GPT-5.6 Luna wins on 12/14 benchmarks
GPT-5.6 Luna wins 12 of 14 shared benchmarks. Leads in agentic · reasoning · knowledge.
Category leads
agentic·GPT-5.6 Lunareasoning·GPT-5.6 Lunaknowledge·GPT-5.6 Lunageneral·GPT-5.6 Lunamath·GPT-5.6 Lunacoding·GPT-5.6 Luna
Hype vs Reality
Attention vs performance
Gemini 2.5 Pro
#126 by perf·no signal
GPT-5.6 Luna
#99 by perf·#9 by attention
Best value
GPT-5.6 Luna
8.5x better value than Gemini 2.5 Pro
Gemini 2.5 Pro
9.0 pts/$
$5.63/M
GPT-5.6 Luna
76.9 pts/$
$0.70/M
Vendor risk
Who is behind the model
Google DeepMind
$3.61T·Tier 1
OpenAI
$840.0B·Tier 1
Head to head
14 benchmarks · 2 models
Gemini 2.5 ProGPT-5.6 Luna
APEX-Agents
GPT-5.6 Luna leads by +36.4
APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments.
Gemini 2.5 Pro
6.6
GPT-5.6 Luna
43.0
ARC-AGI
GPT-5.6 Luna leads by +47.0
ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization.
Gemini 2.5 Pro
41.0
GPT-5.6 Luna
88.0
ARC-AGI-2
GPT-5.6 Luna leads by +54.7
ARC-AGI-2 · the second iteration of the Abstraction and Reasoning Corpus, testing novel pattern recognition and abstract reasoning without prior training data.
Gemini 2.5 Pro
4.9
GPT-5.6 Luna
59.5
Balrog
GPT-5.6 Luna leads by +2.3
Balrog · benchmarks AI agents on text-based adventure games, testing language understanding, strategic planning, and long-horizon reasoning.
Gemini 2.5 Pro
43.3
GPT-5.6 Luna
45.6
Chess Puzzles
GPT-5.6 Luna leads by +24.2
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
Gemini 2.5 Pro
15.8
GPT-5.6 Luna
40.0
Dtbench
GPT-5.6 Luna leads by +11.1
Gemini 2.5 Pro
70.7
GPT-5.6 Luna
81.8
FrontierMath-Tier-4-v2-Private
GPT-5.6 Luna leads by +61.0
Gemini 2.5 Pro
0.0
GPT-5.6 Luna
61.0
FrontierMath-Tiers-1-3-v2-Private
GPT-5.6 Luna leads by +60.6
Gemini 2.5 Pro
25.9
GPT-5.6 Luna
86.4
GPQA diamond
GPT-5.6 Luna leads by +8.8
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Gemini 2.5 Pro
83.7
GPT-5.6 Luna
92.5
Lmca
GPT-5.6 Luna leads by +16.2
Gemini 2.5 Pro
40.9
GPT-5.6 Luna
57.1
OTIS Mock AIME 2024-2025
GPT-5.6 Luna leads by +13.6
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
Gemini 2.5 Pro
84.7
GPT-5.6 Luna
98.3
SimpleBench
Gemini 2.5 Pro leads by +18.7
SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking.
Gemini 2.5 Pro
54.9
GPT-5.6 Luna
36.2
SimpleQA Verified
Gemini 2.5 Pro leads by +15.0
SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information.
Gemini 2.5 Pro
56.0
GPT-5.6 Luna
41.0
WeirdML
GPT-5.6 Luna leads by +6.8
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
Gemini 2.5 Pro
54.0
GPT-5.6 Luna
60.9
Full benchmark table
| Benchmark | Gemini 2.5 Pro | GPT-5.6 Luna |
|---|---|---|
APEX-Agents APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments. | 6.6 | 43.0 |
ARC-AGI ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization. | 41.0 | 88.0 |
ARC-AGI-2 ARC-AGI-2 · the second iteration of the Abstraction and Reasoning Corpus, testing novel pattern recognition and abstract reasoning without prior training data. | 4.9 | 59.5 |
Balrog Balrog · benchmarks AI agents on text-based adventure games, testing language understanding, strategic planning, and long-horizon reasoning. | 43.3 | 45.6 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 15.8 | 40.0 |
Dtbench | 70.7 | 81.8 |
FrontierMath-Tier-4-v2-Private | 0.0 | 61.0 |
FrontierMath-Tiers-1-3-v2-Private | 25.9 | 86.4 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 83.7 | 92.5 |
Lmca | 40.9 | 57.1 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 84.7 | 98.3 |
SimpleBench SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking. | 54.9 | 36.2 |
SimpleQA Verified SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information. | 56.0 | 41.0 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 54.0 | 60.9 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $1.25 | $10.00 | 1.0M tokens (~524 books) | $34.38 | |
| $0.20 | $1.20 | 1.1M tokens (~525 books) | $4.50 |