Windsurf: Gemini 3.1 Pro Now Available
Windsurf added Gemini 3.1 Pro to its model lineup on February 19, 2026, making Google's latest frontier model accessible directly within the IDE. The model is offered with limited-time promotional pricing for self-serve users and supports both Low Thinking and High Thinking modes, giving developers fine-grained control over reasoning depth. Gemini 3.1 Pro delivers more than double the reasoning performance of its predecessor and achieves 77.1% on ARC-AGI-2, making it a significant upgrade for complex engineering tasks.
Sources & Mentions
5 external resources covering this update
Gemini 3.1 Pro is now available in Windsurf
X (Twitter)
Google launches Gemini 3.1 Pro, retaking AI crown with 2X+ reasoning performance boost
VentureBeat
Google's new Gemini Pro model has record benchmark scores β again
TechCrunch
Gemini 3.1 Pro
Hacker News
Google Gemini 3.1 Pro first impressions: a Deep Think Mini with adjustable reasoning on demand
VentureBeat
Gemini 3.1 Pro Arrives in Windsurf
Windsurf expanded its AI model selection on February 19, 2026 with the addition of Gemini 3.1 Pro β Google's latest frontier model and one of the strongest reasoning models available as of its release date. The integration gives Windsurf users direct access to the model inside the IDE, with no need to switch between tools or manage API keys manually.
Promotional Pricing for Self-Serve Users
Windsurf introduced Gemini 3.1 Pro with limited-time promotional credit pricing:
- Low Thinking: 0.5x credits
- High Thinking: 1x credits
The Low and High Thinking modes map to the model's adjustable reasoning depth, a key differentiator of Gemini 3.1 Pro. Developers can opt for faster, lighter responses when exploring ideas, or dial up the reasoning budget for tasks that require multi-step planning, architecture decisions, or research synthesis.
About Gemini 3.1 Pro
Released by Google on February 19, 2026, Gemini 3.1 Pro represents a substantial leap over Gemini 3 Pro. The model scores 77.1% on ARC-AGI-2 β more than double the reasoning performance of its predecessor β and 80.6% on SWE-Bench Verified, placing it among the top-performing models for software engineering benchmarks. Google maintained the same pricing structure as Gemini 3 Pro ($2 per million input tokens via API), meaning the performance upgrade comes at no additional cost.
The model's three-tier thinking system is its most notable architectural feature, allowing users and tools to control how much compute the model invests in each response. This makes it particularly well-suited for agentic coding workflows where the appropriate reasoning depth varies depending on the task β from quick autocomplete suggestions to complex refactoring sessions.
Availability
Gemini 3.1 Pro is available to all Windsurf self-serve users immediately upon the February 19 release.