Gemini 2.5 Pro vs Claude Sonnet 4.6

Both are in the $1-5 per million range and compete for the "standard tier production workhorse" slot.

Gemini 2.5 Pro — $1.25 in / $10 out / $0.31 cached. 2M context. Strong on RAG, long-doc, multimodal. Weaker on multi-tool agent loops.

Claude Sonnet 4.6 — $3 in / $15 out / $0.30 cached. 200K context. Best-in-tier tool use, agentic coding, instruction following.

Pure cost: Gemini wins. A 10K-request chatbot (600 in / 200 out / 60% cache) is $22 on Gemini vs $56 on Sonnet.

Coding agents: Sonnet wins. Tool loops hold together better. SWE-bench gap (~60% vs ~50%) is meaningful.

Long documents: Gemini wins outright past 200K tokens; Sonnet doesn't fit.

Multimodal: Gemini is more developed on images, audio, and video; Sonnet is competitive on images.

Practical verdict: Default to Gemini 2.5 Pro for the cost. Upgrade to Sonnet 4.6 when you actually need tool reliability or coding quality that Gemini keeps dropping.