Automated page. Written and translated automatically by AI and checked against the sources by software; no human editor. SI-IQ is an index built from public benchmark results, not an IQ test. How it works
Gemini 4 Argon vs GPT-6.1 Sol: which is smarter?
Gemini 4 Argon has an SI-IQ of 140 and GPT-6.1 Sol has 148. The table compares them area by area using only independent benchmark results.
| Metric | Gemini 4 Argon | GPT-6.1 Sol |
|---|---|---|
| SI-IQ | 140137–143 | 148144–152 |
| Reasoning | 142 | 150 |
| Math | 138 | 148 |
| Knowledge | 148 | 147 |
| Developer | Google DeepMind | OpenAI |
| Released | Sep 30, 2026 | Sep 29, 2026 |
| Access | Not stated | API access |
| Price in / out (USD per 1M tokens) | $2.00 / $10 | $2.00 / $10 |
| Benchmarks used | 8 | 8 |
Benchmarks every compared model was tested on
Raw scores as published by the evaluators.
| Benchmark | Gemini 4 Argon | GPT-6.1 Sol |
|---|---|---|
| Chess Puzzles | 59.0% | 68.0% |
| FrontierMath Tier 4 | 78.0% | 100.0% |
| FrontierMath Tiers 1–3 | 86.0% | 93.7% |
| Furniture Assembly | 72.5% | 80.0% |
| GPQA Diamond | 96.4% | 95.4% |
| GPQA Diamond (OpenRouter run) | 96.5% | 94.4% |
| Mystery Game Puzzles | 64.0% | 80.0% |
| OTIS Mock AIME 2024–2025 | 100.0% | 100.0% |
| SimpleQA Verified | 76.0% | 73.9% |
| τ²-bench Airline (OpenRouter run) | 78.7% | 68.2% |
On SI-IQ, GPT-6.1 Sol is ahead by 8 points, and the 90% intervals do not overlap.
Data sources: Arena leaderboard dataset on Hugging Face (CC BY 4.0); Epoch AI, ‘Capabilities & benchmarking’ (CC BY 4.0); API list prices via LiteLLM's price table (MIT); each price links to the developer's pricing page; Source: OpenRouter (openrouter.ai), licensed under CC BY 4.0: reuse and republish with attribution to OpenRouter..