OpenAI's and Google's top models are the widest-spread flagship pair on price right now. Two scheduled or conditional changes — Argon's intro rate ending and Astra's long-context band — move the ratio a lot.
Gemini 4 Argon is 5× cheaper than GPT-6 Astra on input and output at today's intro rate. When Argon doubles on December 1 the gap narrows to 2.5× on standard prompts — but on requests above 272K tokens, where Astra's rate goes up and Argon's does not, it widens again.
| GPT-6 Astra | Gemini 4 Argon | |
|---|---|---|
| Provider | OpenAI | |
| Tier | Flagship | Flagship |
| Input / 1M tokens | $10.00 | $2.00 |
| Output / 1M tokens | $50.00 | $10.00 |
| Cache hit / 1M | $1.00 | $0.10 |
| Cache write / 1M | $12.50 | — |
| Context window | 1M | 1M |
| Notes | New top flagship (Sep 3, 2026); rolling out to Plus/Pro/Business/Enterprise; >272K tokens: $20/$75, cache $2 | New top frontier model, generally available Oct 13, 2026 for AI Pro/Ultra subscribers and via the API (model id gemini-4-argon); long-decode generation up to 1M tokens; intro price thru Nov 30, 2026, then $4/$20 (cache read $0.20); not available on the free tier |
Gemini 4 Argon is 5.0× cheaper on input tokens ($2.00 vs $10.00 per 1M). Gemini 4 Argon is 5.0× cheaper on output tokens ($10.00 vs $50.00 per 1M).
| Workload | GPT-6 Astra | Gemini 4 Argon | Difference |
|---|---|---|---|
| Support chatbot 10,000 conversations × (2,000 input + 400 output tokens) | $400 | $80.00 | Gemini 4 Argon saves $320 |
| Document Q&A 1,000 questions × (30,000 input + 800 output tokens) | $340 | $68.00 | Gemini 4 Argon saves $272 |
| Coding agent session 4M input tokens (90% cache hits) + 150,000 output tokens | $15.10 | $2.66 | Gemini 4 Argon saves $12.44 |
Standard rates per 1M tokens from data on this site. The coding-agent row assumes 90% of input is read from the prompt cache (full input price where a model has no cache rate). Cache-write fees, batch discounts and long-context surcharges listed in the notes are not included. Updated November 2, 2026.
your stack is built on OpenAI's API, you need OpenAI-specific tooling, or your evals show Astra finishing tasks in fewer attempts — and your prompts mostly stay under 272K tokens.
cost per token is a deciding factor, you send very long prompts, or you want caching that works without code changes (Google applies Argon's cache discount implicitly).
Through November 30 Argon costs $2.00/$10.00. From December 1 Google's list price is double that, still well below Astra's $10.00/$50.00. In the other direction, Astra's requests over 272K tokens cost 2× on input and 1.5× on output, while Google lists no long-context tier for Argon. A team that routinely sends 300K-token prompts sees a bigger gap after December than a chatbot team does today.
Astra charges $12.50 per 1M tokens to write a 30-minute cache (a 12-hour option costs more) and $1.00 per cached read. Argon's caching is implicit: no write fee is listed, and repeated prefixes are billed at $0.10. For agent loops that re-send a large context, Argon also avoids the operational work of managing cache lifetimes.
Both providers halve prices for Batch jobs, and OpenAI also has a half-price Flex tier for latency-tolerant traffic. If you want to stay with OpenAI but cut cost, GPT-6.1 Sol sits below Astra at $2.00/$10.00 — roughly Argon's intro rate. Argon has no free API tier, so budget for testing.
About 5× on input and output at the intro rate that runs through November 30, 2026, and about 2.5× after Argon's list price doubles on December 1.
GPT-6 Astra does: requests over 272K tokens are billed at a higher input, output and cache-hit rate. Google lists no long-context tier for Gemini 4 Argon.
Gemini 4 Argon, at $0.10 per 1M tokens versus $1.00 on Astra, with no listed cache-write fee.
How we choose: prices come from each vendor's official pricing page and are re-checked every week. We take no affiliate fees, so rankings aren't for sale.