Claude Sonnet 5.5 vs Gemini 4 Argon
Claude Sonnet 5.5 lists at $2 / $10 per 1M tokens. Gemini 4 Argon shows the same $2 / $10 today, but only as an introductory rate.
Our verdict
The sticker prices match today, but Argon's rate is promotional through 30 Nov 2026 and rises to $4 / $20, while Claude Sonnet 5.5 stays put. Artificial Analysis scores Claude Sonnet 5.5 higher (56 against 53, 1 Oct 2026 capture), though it found Sonnet very wordy (about $7.60 per index task). Argon's headline is Google's own 77.9% on DeepSWE, a vendor-reported coding result. For a budget you can plan around, choose Claude Sonnet 5.5; use the promo to test Gemini 4 Argon on your hardest coding job.
Benchmarks measure test tasks, not your business. Run a real job through both models before you commit.
Side by side
| Claude Sonnet 5.5 | Gemini 4 Argon | |
|---|---|---|
| Lab | Anthropic | |
| List price per 1M tokens (input / output) | $2 / $10 | $2 / $10 |
| Scheduled price | - | From Dec 1, 2026: $4 / $20 |
| Price notes | The full 1M-token window bills at the standard rate; cache reads $0.10 per 1M. | Introductory rate through 30 Nov 2026. |
| Example job at list price | $4 | $4 |
| Context window | 1,000,000 tokens | Not published |
| Released | ||
| Open weights | No | No |
| Artificial Analysis Intelligence Index | 56 (Oct 1, 2026 capture) | 53 (Oct 1, 2026 capture)at high effort |
1M tokens in and 200K tokens out. List price only: models write different amounts, so a finished job can cost more or less. (1,000,000 / 200,000)
Common questions
Do Claude Sonnet 5.5 and Gemini 4 Argon cost the same?
Today yes: $2 in and $10 out per 1M tokens. Argon's price is introductory through 30 November 2026 and becomes $4 and $20 from 1 December 2026.
Which scores higher on independent benchmarks?
Artificial Analysis scored Sonnet 5.5 56 and Gemini 4 Argon 53 (1 Oct 2026 capture).
Which is better for coding?
Google reports 77.9% on DeepSWE for Argon and Anthropic reports 70.6% on Terminal-Bench 4.0 for Sonnet 5.5, but those are different benchmarks and vendor-reported. Test both on your own repository.
Sources
More model comparisons
All model comparisonsSee every model on the AI model tracker
Not sure which model your business needs?
Dapols picks the models and tools for your budget and shows the bill before you spend.