Claude Sonnet 4.6 and Gemini 3.1 Pro are both premium models at similar prices but with different strengths. Claude leads on coding (79.6% SWE-bench) and writing quality. Gemini 3.1 Pro wins on research depth and is 33% cheaper ($2 vs $3/1M input); both read about 1M tokens. For developers and content teams, Claude is the stronger daily driver — today that means Claude Sonnet 5.5. For research and cost-sensitive workloads, Gemini is the smarter pick.
AnthropicPremium
Claude Sonnet 4.6
Earlier Sonnet — replaced by Sonnet 5 and then Sonnet 5.5.
VS
GooglePremium
Gemini 3.1 Pro
Google's Pro model — 1M context and native multimodal input at $2/$12.
At a glance
Claude Sonnet 4.6
Gemini 3.1 Pro
Input cost / 1M tokens
$$3.00/1M
$$2.00/1M
Output cost / 1M tokens
$$15.00/1M
$$12.00/1M
Context window
1M tokens
1M tokens
Speed
Balanced
Balanced
Price tier
Premium
Premium
Benchmarks
SWE-bench (coding)
79.6%
80.6%
Arena Elo
1,340
1,380
MMLU
88.3%
90%
How they compare
Which model wins for each use case — and why.
CodingClaude Sonnet 4.6 wins
Claude Sonnet 4.6 scores 79.6% on SWE-bench and is the default in Cursor and Windsurf — the leading AI code editors. Gemini handles code but trails on benchmarks.
WritingClaude Sonnet 4.6 wins
Claude Sonnet 4.6 leads the writing category and consistently produces cleaner, more natural prose with better tone control than Gemini 3.1 Pro.
ResearchGemini 3.1 Pro wins
Gemini 3.1 Pro scored 77.1% on ARC-AGI-2 at launch and leads our research measure, at a lower price.
Context WindowTie
Both support about 1M tokens, so neither has a capacity advantage.
PriceGemini 3.1 Pro wins
Gemini 3.1 Pro costs $2/1M input vs Claude Sonnet 4.6's $3/1M — 33% cheaper. Output is also cheaper at $12 vs $15/1M.
Which should you pick?
Pick Claude Sonnet 4.6 if…
Coding is your primary use case — Claude leads on SWE-bench and AI code editors
Writing quality, tone, and content polish matter
You use Cursor or Windsurf, which default to Claude
You want a strong all-rounder for daily developer and content work
Against Gemini 3.1 Pro it costs about 22% more per token.
The default model powering Cursor and Windsurf. 79.6% SWE-bench, 1M context window, and best-in-tier writing quality — all at $3/1M input.
Input
$3.00/1M
Output
$15.00/1M
Context
1M tokens
Speed
Balanced
What people actually use it for
Daily coding in Cursor — debugging, refactoring, and feature implementation
Drafting polished client reports, strategy memos, and long-form editorial content
Answering research questions with up to 1M tokens of document context
Where it wins
Strong coding quality with 1M context at $3/1M input
Default model in Cursor and Windsurf, the two most popular AI coding editors
Best writing quality in its price tier — tone, long-form clarity, editorial polish
Where it falls down
Replaced by Claude Sonnet 5 and then Sonnet 5.5, which is stronger and cheaper
GPT-5.5 or GPT-5.4 are better picks when OpenAI computer-use workflows are the priority
Skip it if
You specifically need desktop-control capabilities (GPT-5.5/GPT-5.4) or the absolute highest coding ceiling (Opus 4.7).
Our verdict
The best all-around model for most developers and writers. Strong SWE-bench, excellent writing, 1M context — all at $3/1M input. Hard to beat as a daily driver.
Against Claude Sonnet 4.6 it costs about 22% less per token.
Google's Pro-tier Gemini 3 model, released in preview on February 19, 2026, with a 1M-token context window, Deep Think reasoning and $2/$12 pricing.
Input
$2.00/1M
Output
$12.00/1M
Context
1M tokens
Speed
Balanced
What people actually use it for
Analyzing long contracts, codebases, or research papers in a single 1M-token prompt
Due diligence synthesis across large sets of financial documents or legal agreements
Multi-step reasoning across dense technical specifications with Deep Think mode
Where it wins
1M-token context window with native image, audio, video and PDF input
Scored 77.1% on ARC-AGI-2 at its February 2026 launch
$2/$12 per 1M tokens — cheaper output than most premium models
Where it falls down
Slower than Flash for everyday lightweight tasks
Claude Sonnet 5.5 is better for writing quality, at a lower $2/$10
Skip it if
Your primary use case is writing quality or agentic coding — Claude wins both.
Our verdict
A capable research model and still Google's current Pro tier, but no longer the long-context leader: its 1M-token window matches Claude Opus 5.5 and is slightly below the GPT-6 family's 1.05M. Choose it for Google Workspace integration and multimodal input; for most research, Claude Opus 5.5 or GPT-6.1 Sol are stronger at a similar price.
Claude Sonnet 4.6 is better for coding and writing. Gemini 3.1 Pro is better for research, large-context work, and cost efficiency. For everyday developer use, Claude wins. For large-document analysis, Gemini wins.
Which is cheaper?
Gemini 3.1 Pro costs $2/$12 per 1M tokens vs Claude Sonnet 4.6's $3/$15. Gemini is 33% cheaper on input and 20% cheaper on output.
Which has a bigger context window?
Neither — both are about 1M tokens. Earlier versions of this page said Gemini 3.1 Pro had 2M; Google's documentation lists 1M.