Compare · ModelsLive · 2 picked · head to head
Gemini 3.8 Flash vs GPT-6.1 Sol
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
GPT-6.1 Sol wins on 16/20 benchmarks
GPT-6.1 Sol wins 16 of 20 shared benchmarks. Leads in speed · knowledge · coding.
Category leads
speed·GPT-6.1 Solagentic·Gemini 3.8 Flashreasoning·Gemini 3.8 Flashknowledge·GPT-6.1 Solcoding·GPT-6.1 Solmath·GPT-6.1 Solgeneral·GPT-6.1 Sol
Hype vs Reality
Attention vs performance
Gemini 3.8 Flash
#53 by perf·#6 by attention
GPT-6.1 Sol
#6 by perf·#5 by attention
Best value
Gemini 3.8 Flash
2.1x better value than GPT-6.1 Sol
Gemini 3.8 Flash
26.5 pts/$
$2.25/M
GPT-6.1 Sol
12.9 pts/$
$6.00/M
Vendor risk
Who is behind the model
Google DeepMind
$4.27T·Tier 1
OpenAI
$840.0B·Tier 1
Head to head
20 benchmarks · 2 models
Gemini 3.8 FlashGPT-6.1 Sol
Artificial Analysis · CritPt
GPT-6.1 Sol leads by +13.4
Gemini 3.8 Flash
18.3
GPT-6.1 Sol
31.7
Artificial Analysis · GDPval
GPT-6.1 Sol leads by +7.9
Gemini 3.8 Flash
46.8
GPT-6.1 Sol
54.7
Artificial Analysis · Humanity's Last Exam
GPT-6.1 Sol leads by +5.1
Gemini 3.8 Flash
47.8
GPT-6.1 Sol
52.9
Artificial Analysis · Long Context Reasoning
GPT-6.1 Sol leads by +1.7
Gemini 3.8 Flash
81.3
GPT-6.1 Sol
83.0
Artificial Analysis · MMMU Pro
GPT-6.1 Sol leads by +0.4
Gemini 3.8 Flash
85.6
GPT-6.1 Sol
86.0
Artificial Analysis · Quality Index
GPT-6.1 Sol leads by +10.9
Gemini 3.8 Flash
40.9
GPT-6.1 Sol
51.8
Artificial Analysis · SciCode
Gemini 3.8 Flash leads by +2.4
Gemini 3.8 Flash
56.6
GPT-6.1 Sol
54.2
APEX-Agents
Gemini 3.8 Flash leads by +4.3
APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments.
Gemini 3.8 Flash
64.3
GPT-6.1 Sol
60.0
ARC-AGI
ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization.
Gemini 3.8 Flash
98.5
GPT-6.1 Sol
98.5
ARC-AGI-2
GPT-6.1 Sol leads by +5.0
ARC-AGI-2 · the second iteration of the Abstraction and Reasoning Corpus, testing novel pattern recognition and abstract reasoning without prior training data.
Gemini 3.8 Flash
89.2
GPT-6.1 Sol
94.2
Chess Puzzles
GPT-6.1 Sol leads by +10.5
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
Gemini 3.8 Flash
55.8
GPT-6.1 Sol
66.3
Frontiercode
GPT-6.1 Sol leads by +9.0
Gemini 3.8 Flash
41.2
GPT-6.1 Sol
50.2
FrontierMath-Tier-4-v2-Private
GPT-6.1 Sol leads by +78.0
Gemini 3.8 Flash
21.9
GPT-6.1 Sol
100.0
FrontierMath-Tiers-1-3-v2-Private
GPT-6.1 Sol leads by +25.3
Gemini 3.8 Flash
68.4
GPT-6.1 Sol
93.7
Furniture Assembly
GPT-6.1 Sol leads by +69.1
Gemini 3.8 Flash
2.4
GPT-6.1 Sol
71.4
GPQA diamond
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Gemini 3.8 Flash
93.9
GPT-6.1 Sol
93.9
Mystery Game Puzzles
GPT-6.1 Sol leads by +36.4
Gemini 3.8 Flash
41.6
GPT-6.1 Sol
78.0
OTIS Mock AIME 2024-2025
GPT-6.1 Sol leads by +1.1
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
Gemini 3.8 Flash
98.9
GPT-6.1 Sol
100.0
Proofbench
GPT-6.1 Sol leads by +51.0
Gemini 3.8 Flash
48.0
GPT-6.1 Sol
99.0
SimpleQA Verified
GPT-6.1 Sol leads by +4.2
SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information.
Gemini 3.8 Flash
69.7
GPT-6.1 Sol
73.9
Full benchmark table
| Benchmark | Gemini 3.8 Flash | GPT-6.1 Sol |
|---|---|---|
Artificial Analysis · CritPt | 18.3 | 31.7 |
Artificial Analysis · GDPval | 46.8 | 54.7 |
Artificial Analysis · Humanity's Last Exam | 47.8 | 52.9 |
Artificial Analysis · Long Context Reasoning | 81.3 | 83.0 |
Artificial Analysis · MMMU Pro | 85.6 | 86.0 |
Artificial Analysis · Quality Index | 40.9 | 51.8 |
Artificial Analysis · SciCode | 56.6 | 54.2 |
APEX-Agents APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments. | 64.3 | 60.0 |
ARC-AGI ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization. | 98.5 | 98.5 |
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. | 89.2 | 94.2 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 55.8 | 66.3 |
Frontiercode | 41.2 | 50.2 |
FrontierMath-Tier-4-v2-Private | 21.9 | 100.0 |
FrontierMath-Tiers-1-3-v2-Private | 68.4 | 93.7 |
Furniture Assembly | 2.4 | 71.4 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 93.9 | 93.9 |
Mystery Game Puzzles | 41.6 | 78.0 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 98.9 | 100.0 |
Proofbench | 48.0 | 99.0 |
SimpleQA Verified SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information. | 69.7 | 73.9 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $0.75 | $3.75 | 1.0M tokens (~524 books) | $15.00 | |
| $2.00 | $10.00 | 1.1M tokens (~525 books) | $40.00 |