The llm-gemini 0.35 release adds support for Google's new Gemini 4 Argon model. Gemini 4 Argon is available today via the Gemini API and to customers on Gemini's $100/month subscription plan.
Illustrations and a cost example
The developer shared three SVG illustrations labeled “low”, “medium” and “high” (pelican-themed), noting they are all excellent. The “high” illustration cost 17.166 cents to generate (0.17166 USD).
Launch pricing and upcoming change
At launch, Gemini 4 Argon is priced at $2 per million input tokens and $10 per million output tokens. That matches the price of GPT-6 Sol, is half the price of Claude Opus 5.5, and one-fifth the price of Claude Fable 5.1 and GPT-6 Astra.
However, on December 1 the Argon price is scheduled to double, which will bring it in line with the Opus 5.5 pricing.
Why raw token prices aren't enough
The author emphasizes that per-token prices alone are increasingly misleading: the important metric is how many tokens a model consumes to accomplish a task, especially when used inside agentic systems. A practical way to assess that is to look at benchmark costs published by third parties.
Benchmark cost data
Artificial Analysis publishes the total cost to run their benchmark with each model. According to their data, running the benchmark on Argon cost $2,407, which places Argon close to GLM-5.3 and substantially cheaper than the latest models from OpenAI, Anthropic, and Meta. Even after the planned price doubling, Argon would still be slightly cheaper than GPT-6 Astra based on those benchmark costs.
Note on chart editing
The author notes they used ChatGPT to edit the chart image for presentation, and states that they verified the underlying data; the edit was to prevent the Gemini 4 Argon info box from overlapping other numbers.
Summary
With llm-gemini 0.35, developers gain official support for Gemini 4 Argon. The model's introductory pricing looks competitive, though the scheduled December price increase will alter cost comparisons; current benchmark-based cost figures suggest Argon is cost-effective relative to several major models.
