Compare · ModelsLive · 2 picked · head to head
o1-preview vs Gemini 2.0 Flash
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
o1-preview wins on 4/6 benchmarks
o1-preview wins 4 of 6 shared benchmarks. Leads in arena · reasoning · coding.
Category leads
arena·o1-previewknowledge·Gemini 2.0 Flashmath·Gemini 2.0 Flashreasoning·o1-previewcoding·o1-preview
Hype vs Reality
Attention vs performance
o1-preview
#136 by perf·no signal
Gemini 2.0 Flash
#101 by perf·no signal
Vendor risk
Who is behind the model
OpenAI
$840.0B·Tier 1
Google DeepMind
$4.00T·Tier 1
Head to head
6 benchmarks · 2 models
o1-previewGemini 2.0 Flash
Chatbot Arena Elo · Overall
o1-preview leads by +27.7
o1-preview
1387.7
Gemini 2.0 Flash
1360.0
GPQA diamond
Gemini 2.0 Flash leads by +18.4
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
o1-preview
33.8
Gemini 2.0 Flash
52.2
MATH level 5
Gemini 2.0 Flash leads by +0.5
MATH Level 5 · the hardest tier of the MATH benchmark, featuring competition-level problems from AMC, AIME, and Olympiad-style mathematics.
o1-preview
81.7
Gemini 2.0 Flash
82.2
OTIS Mock AIME 2024-2025
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
o1-preview
31.0
Gemini 2.0 Flash
31.0
SimpleBench
o1-preview leads by +12.7
SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking.
o1-preview
30.0
Gemini 2.0 Flash
17.3
WeirdML
o1-preview leads by +21.8
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
o1-preview
47.6
Gemini 2.0 Flash
25.8
Full benchmark table
| Benchmark | o1-preview | Gemini 2.0 Flash |
|---|---|---|
Chatbot Arena Elo · Overall | 1387.7 | 1360.0 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 33.8 | 52.2 |
MATH level 5 MATH Level 5 · the hardest tier of the MATH benchmark, featuring competition-level problems from AMC, AIME, and Olympiad-style mathematics. | 81.7 | 82.2 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 31.0 | 31.0 |
SimpleBench SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking. | 30.0 | 17.3 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 47.6 | 25.8 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| — | — | — | — | |
| $0.10 | $0.40 | 1.0M tokens (~500 books) | $1.75 |
People also compared