Cost by workload size
Price comparison
| Feature |  Claude Sonnet 4.6 |  Gemini 3.1 Pro |
|---|---|---|
| Provider | Anthropic | |
| Input / M tokens | $3 | $2 |
| Output / M tokens | $15 | $12 |
| Cached input read | $0.30 | — |
| Context window | 1M | 1.0M |
| Batch discount | — | −50% |
| Vision | Yes | Yes |
| Audio | — | Yes |
| Reasoning tokens | — | Yes |
What your workload actually costs
Prices per token hide the real picture. Here are three common workloads computed with AITokenCalculator's own engine (medium reasoning effort, no batch, no cache):
| Workload | Claude Sonnet 4.6 | Gemini 3.1 Pro | Cheaper |
|---|---|---|---|
| 1,000 in / 300 out | $0.0075 | $0.0092 | Claude Sonnet 4.6 |
| 10,000 in / 2,000 out | $0.06 | $0.068 | Claude Sonnet 4.6 |
| 100,000 in / 20,000 out | $0.6 | $0.68 | Claude Sonnet 4.6 |
| Monthly @ 1,000 req/day (standard) | $1,826.25 | $2,069.75 | Claude Sonnet 4.6 |
Analysis
On input price, Gemini 3.1 Pro wins at $2/M — roughly 1.5× cheaper than its rival. Output tells a sharper story: Gemini 3.1 Pro charges $12/M , about 1.3× less than Claude Sonnet 4.6 — and output is where chatty and agentic workloads bleed money.
Context capacity differs too: Gemini 3.1 Pro fits 1.0M tokens (~786,432 words) versus 1M. If you feed whole documents or codebases, that gap decides feasibility before price even matters.
Long-context specialists. Gemini Pro brings the bigger window for whole-repository or multi-file analysis; Claude Sonnet counters with stronger instruction adherence and code editing. If your prompts exceed 200K tokens the choice makes itself.
Numbers shift as providers reprice — check the full AI model pricing table, then run your own prompt through the calculator preloaded with Claude Sonnet 4.6 or with Gemini 3.1 Pro.