Gemini 4 Argon vs Claude Opus 5.5: The Full Table and Which Model to Use
Google says its new model beats Claude Opus 5.5 on 14 of 18 tests. The full table, the four tests Opus still wins, what Argon costs next to Claude, and a cheat sheet for picking a model by job
On October 13, Google released Gemini 4 Argon. On the benchmark table in Google's own model card it scores higher than Claude Opus 5.5 on 14 of the 18 tests where Opus has a score, and lower on four. Here is the whole table, what it means for someone who uses Claude, and a cheat sheet.
One caution first: every number below comes from Google's table. Google does not say how it got the Claude scores.
The full table (Google's own numbers)
| Test | Gemini 4 Argon | GPT-6 Astra | Claude Fable 5.1 | Claude Opus 5.5 |
|---|---|---|---|---|
| Vals Index | 68.9% | 63.1% | 65.8% | 67.0% |
| AutomationBench | 51.3% | 41.4% | 31.4% | 42.5% |
| Vals Finance Agent v2 | 65.4% | 53.5% | 58.9% | 58.6% |
| Harvey's Legal Agent Benchmark | 19.6% | 5.4% | 6.7% | 3.8% |
| DeepSWE v1.1 | 77.9% | 74.1% | 67.4% | 74.2% |
| FrontierSWE v2 | 55.0% | 65.5% | 56.3% | 62.3% |
| Vibe Code Bench | 91.9% | 89.6% | 90.3% | 90.3% |
| Terminal-bench 4.0 | 57.4% | 58.2% | 57.9% | 66.4% |
| PostTrainBench | 45.3% | 44.3% | 40.2% | 49.3% |
| Terminal-Bench Science 0.1 | 57.6% | 68.1% | 52.6% | 63.3% |
| LABBench 2 | 88.8% | 85.4% | 68.6% | 73.1% |
| RiemannBench | 76.0% | 72.0% | 65.6% | 69.6% |
| GraphWalks, up to 128k (BFS F1) | 99.7% | 98.7% | 91.4% | 90.6% |
| GraphWalks, 256k to 1M (BFS F1) | 84.2% | 71.8% | 65.0% | 66.8% |
| Agent's Last Exam | 39.5% | 34.2% | - | 38.2% |
| OSWorld-2.0 (offline subset) | 69.2% | 72.6% | - | - |
| Chartography | 71.6% | 71.0% | 46.2% | 66.3% |
| LVBench | 91.7% | 87.5% | 79.7% | 83.7% |
| CWE-bench v1 | 68.0% | 68.0% | 58.0% | 67.0% |
A dash means Google shows no score for that model. Counting the 18 rows where Opus 5.5 has a score, Argon is ahead on 14 and Opus is ahead on 4.
Where Claude still wins
- Terminal-bench 4.0: Opus 5.5 66.4%, Argon 57.4%. This is the one closest to a coding agent working in a terminal.
- FrontierSWE v2: Opus 5.5 62.3%, Argon 55.0%.
- PostTrainBench: Opus 5.5 49.3%, Argon 45.3%.
- Terminal-Bench Science 0.1: Opus 5.5 63.3%, Argon 57.6%.
- Argon wins DeepSWE v1.1 (77.9% vs 74.2%) and Vibe Code Bench (91.9% vs 90.3%), so coding is split, not one-sided.
Where Argon wins
- Legal and finance agents: Harvey's Legal Agent Benchmark 19.6% vs 3.8%, Vals Finance Agent v2 65.4% vs 58.6%.
- Office-style automation: AutomationBench 51.3% vs 42.5%. Overall Vals Index 68.9% vs 67.0%.
- Long context: GraphWalks from 256k to 1M tokens 84.2% vs 66.8%. Argon's window is 1,048,576 tokens.
- Images and video: Chartography 71.6% vs 66.3%, LVBench 91.7% vs 83.7%.
Free, for your email
Get the full setup steps + the copy-paste setup prompt
The rest of this guide (5 more sections, 1 copy-paste prompt) opens right here. The setup prompt, one paste into Claude that sets it all up, comes by email only. Use your real address.
Unlocks: Cheat sheet: which model for which job · What it costs · How to try Argon · The catch · Sources
50,000+ follow where these guides come from. No spam, reply “unsubscribe” any time.