Neither wins outright: Anthropic: Claude Sonnet 4 is the pick when you need to fit long inputs in one request, Google: Nano Banana Pro (Gemini 3 Pro Image) when cost matters at scale.
| Spec | Anthropic: Claude Sonnet 4 | Google: Nano Banana Pro (Gemini 3 Pro Image) |
|---|---|---|
| Provider | Anthropic | |
| Input price (per 1M tokens) | $3.00 | $2.00 |
| Output price (per 1M tokens) | $15.00 | $12.00 |
| Blended price (3:1 input/output) | $6.00 | $4.50 |
| Cheaper than … of all models we track | 8% | 12% |
| Context window | 200K tokens (~300 pages) | 131K tokens (~197 pages) |
| Larger context than … of all models | 37% | 22% |
| Max output per response | 64K tokens (~48,000 words) | 33K tokens (~24,600 words) |
| Input types | image, text, file | image, text |
| Output types | text | image, text |
| Free variant | No | No |
| Listed since | 2025-05-22 | 2026-06-18 |
Monthly cost at list price for five common workloads. Token counts per unit are stated so you can scale them to your own traffic. A workload that does not fit a model's context window is marked rather than priced.
| Workload (per month) | Anthropic: Claude Sonnet 4 | Google: Nano Banana Pro (Gemini 3 Pro Image) | Difference |
|---|---|---|---|
| Customer-support chatbot 50,000 conversations — each a 6-turn conversation (1,500 in / 400 out tokens) | $525 | $390 | Google: Nano Banana Pro (Gemini 3 Pro Image) saves $135 |
| RAG search over documents 20,000 questions — each a question answered from ~8 retrieved passages (6,000 in / 500 out tokens) | $510 | $360 | Google: Nano Banana Pro (Gemini 3 Pro Image) saves $150 |
| Coding / tool-using agent 2,000 tasks — each a multi-step task with tool calls and file context (60,000 in / 6,000 out tokens) | $540 | $384 | Google: Nano Banana Pro (Gemini 3 Pro Image) saves $156 |
| Batch summarisation 10,000 documents — each a 6,000-word report summarised to a page (8,000 in / 600 out tokens) | $330 | $232 | Google: Nano Banana Pro (Gemini 3 Pro Image) saves $98.00 |
| Long-document analysis 500 documents — each a 150-page contract or codebase read in one pass (100,000 in / 2,000 out tokens) | $165 | $112 | Google: Nano Banana Pro (Gemini 3 Pro Image) saves $53.00 |
| Capability | Anthropic: Claude Sonnet 4 | Google: Nano Banana Pro (Gemini 3 Pro Image) | What it lets you do |
|---|---|---|---|
| Tool / function calling | ✓ Yes | ✓ Yes | call your APIs and run agent loops |
| Structured output (JSON schema) | — | ✓ Yes | return JSON that validates against your schema |
| Reasoning / thinking mode | ✓ Yes | ✓ Yes | spend extra tokens thinking before answering hard problems |
| Image input | ✓ Yes | ✓ Yes | read screenshots, charts and scanned pages |
| Audio input | — | — | take speech or audio directly |
| File / PDF input | ✓ Yes | — | accept documents without your own parsing step |
| Image output | — | ✓ Yes | generate images, not just text |
| Built-in web search | ✓ Yes | ✓ Yes | answer from live web results |
| Seed (reproducible sampling) | — | ✓ Yes | repeat a generation for tests and evals |
| Log-probabilities | — | — | read token confidence for classification and scoring |
From each model's published parameters and input/output types. “—” means not advertised, not proven absent.
Neither wins outright: Anthropic: Claude Sonnet 4 is the pick when you need to fit long inputs in one request, Google: Nano Banana Pro (Gemini 3 Pro Image) when cost matters at scale. In short: Anthropic: Claude Sonnet 4 when you need to fit long inputs in one request, you need file / PDF input, you generate long outputs in one go; Google: Nano Banana Pro (Gemini 3 Pro Image) when cost matters at scale, you need structured output (JSON schema), you need image output, you need seed (reproducible sampling), you want the newer model.
Google: Nano Banana Pro (Gemini 3 Pro Image) is cheaper: $4.50 vs $6.00 per 1M tokens at a typical 3:1 input/output mix, 25% less. For 50,000 support conversations a month that is $390 against $525.
Anthropic: Claude Sonnet 4 costs $3.00 per 1M input tokens and $15.00 per 1M output tokens; Google: Nano Banana Pro (Gemini 3 Pro Image) costs $2.00 input and $12.00 output. Output tokens are billed separately, so long answers shift the gap.
Anthropic: Claude Sonnet 4: 200K tokens, about 300 pages, against 131K (about 197 pages) for Google: Nano Banana Pro (Gemini 3 Pro Image).
Tool calling — Anthropic: Claude Sonnet 4: yes, Google: Nano Banana Pro (Gemini 3 Pro Image): yes. JSON-schema structured output — Anthropic: Claude Sonnet 4: not advertised, Google: Nano Banana Pro (Gemini 3 Pro Image): yes.
Yes, both accept image input.
No free variant is listed for either; both are pay-per-token.
Anthropic: Claude Sonnet 4 can write up to 64K tokens (about 48,000 words) per response; Google: Nano Banana Pro (Gemini 3 Pro Image) up to 33K (about 24,600 words).
Anthropic: Claude Sonnet 4 was listed on 2025-05-22; Google: Nano Banana Pro (Gemini 3 Pro Image) on 2026-06-18.
Claude Sonnet 4 significantly enhances the capabilities of its predecessor, Sonnet 3.7, excelling in both coding and reasoning tasks with improved precision and controllability. Achieving state-of-the-art performance on SWE-bench (72.7%),...
Full Anthropic: Claude Sonnet 4 specs →Nano Banana Pro is Google’s most advanced image-generation and editing model, built on Gemini 3 Pro. It extends the original Nano Banana with significantly improved multimodal reasoning, real-world grounding, and...
Full Google: Nano Banana Pro (Gemini 3 Pro Image) specs →