Gemini 3.6 Flash vs Claude Sonnet 5
These are the two default workhorses of mid-2026, and they optimize for different things. Gemini 3.6 Flash, released July 21 at $1.50 per million input tokens and $7.50 output, now runs $0.75 and $3.75 through December 31, 2026 before reverting to the launch rate, takes text, image, audio, video, and PDF input, and is engineered around token efficiency: Google reports 17 percent fewer output tokens than its predecessor and Artificial Analysis measured time per task falling to 1.3 minutes. Claude Sonnet 5, released June 30, runs $2 per million input and $10 per million output, a launch rate Anthropic made standard when it cancelled a planned September increase, and leads on published coding numbers with vendor-reported 85.2 percent on SWE-bench Verified and 80.4 percent on Terminal-Bench 2.1. Both ship 1 million token context windows. The trade is roughly this: Sonnet 5 is the stronger single-shot coder on the numbers each vendor published, Gemini 3.6 Flash is cheaper per token today, by more than half through the end of 2026, and cheaper still per completed task. Neither set of scores has been independently reproduced on a common harness, so treat the benchmark rows as claims and run your own workload before committing.
Head-to-Head Specs
| Spec | Gemini 3.6 Flash | Claude Sonnet 5 |
|---|---|---|
| Provider | Anthropic | |
| Input Price | $0.75/1M | $2.00/1M |
| Output Price | $3.75/1M | $10.00/1M |
| Context Window | 1.0M | 1M |
| Released | 2026-07 | 2026-06 |
| Capabilities | text, vision, audio, video, tool-use, code, reasoning | text, vision, tool-use, code, reasoning |
Benchmark Scores
| Benchmark | Gemini 3.6 Flash | Claude Sonnet 5 | Winner |
|---|---|---|---|
| SWE-bench | 79.6 | 85.2 | Claude |
| MMLU-Pro | 89.3 | 87.5 | Gemini |
| GPQA Diamond | 93.4 | 88.9 | Gemini |
| FrontierCode v1.1 | 34.4 | 42.7 | Claude |
See the full benchmark leaderboard for all models.
Category Breakdown
$0.75 vs $2.00 through December 31, 2026, and $1.50 vs $2.00 after Gemini reverts
$3.75 vs $10.00 through 2026, and $7.50 vs $10.00 from January 1, 2027
Anthropic reports 85.2 percent on SWE-bench Verified and 80.4 on Terminal-Bench 2.1; Google published no comparable Verified number for 3.6 Flash
3.6 Flash is explicitly tuned to use fewer output tokens and fewer tool calls per task, which compounds with the lower rate
Gemini takes text, image, audio, video, and PDF; Sonnet 5 is text and vision only
Both ship 1 million tokens; Sonnet 5 adds context compaction, Gemini 3.6 Flash allows 64K output
Google reports 83.0 on OSWorld-Verified with computer use as a built-in API tool
Sonnet 5 holds $2/$10 as its standard rate; Gemini 3.6 Flash doubles back to $1.50/$7.50 on January 1, 2027
Sonnet 5 runs on the Claude API, Bedrock, Vertex, and Foundry, and slots into the MCP and Claude Code tooling most agent stacks already target
Choose Gemini 3.6 Flash when:
- ▸High-volume agentic runs where cost per completed task is the binding constraint
- ▸Multimodal pipelines that need audio, video, or PDF input natively
- ▸Computer-use and GUI automation
- ▸Budgets planned through December 31, 2026, while the promotional rate applies
Choose Claude Sonnet 5 when:
- ▸Hard single-shot coding and repository-scale refactors
- ▸Stacks already built on MCP, Claude Code, or Bedrock and Foundry deployments
- ▸Work that benefits from context compaction on very long sessions
- ▸Teams that want one vendor across Sonnet, Opus, and Fable tiers
Frequently Asked Questions
Which is better, Gemini 3.6 Flash or Claude Sonnet 5?
It depends on your use case. Gemini 3.6 Flash from Google excels at high-volume agentic runs where cost per completed task is the binding constraint, while Claude Sonnet 5 from Anthropic is better for hard single-shot coding and repository-scale refactors. See the full comparison above for detailed benchmarks and pricing.
How much does Gemini 3.6 Flash cost compared to Claude Sonnet 5?
Gemini 3.6 Flash costs $0.75 input and $3.75 output per 1M tokens. Claude Sonnet 5 costs $2.00 input and $10.00 output per 1M tokens.
What is the context window difference between Gemini 3.6 Flash and Claude Sonnet 5?
Gemini 3.6 Flash supports 1.0M tokens, while Claude Sonnet 5 supports 1M tokens.