Why Compare All Google Models Together
TokenRate's new Compare Prices grid puts every model's per-token rates, context window, and quality score in a single side-by-side view. The point: stop flipping between provider pricing pages and OpenRouter tabs. You pick a provider dropdown, check the models you want, repeat for each provider, and the grid stacks every pick into one comparison table. Picking between Google's models — flagship, mid-tier, fast — is usually done by reading the provider's pricing page top-to-bottom, which buries the spread. The Compare Prices grid flips that: you check **every Google model** in one dropdown and the grid lays them out side-by-side. Input cost across the lineup spans $0.075 to $1.25 per 1M tokens — a 16.7× spread you can scan in three seconds. This guide walks through the lineup model-by-model with the framing of "what do you give up to step down a tier." Pair this with flagship/balanced/fast/reasoning LLM tiers, Arena AI leaderboard Elo scores explained, and how LLM quality scores are calculated.
The Google Lineup, Top to Bottom
**Gemini 2.5 Pro** — $1.25 input / $10.00 output per 1M, 1M context, quality 78, tier **balanced**. balanced tier is the production-default zone — quality high enough for customer traffic, price low enough to scale. **Gemini 2.5 Flash** — $0.300 input / $2.50 output per 1M, 1M context, quality 68, tier **fast**. fast tier is built for high-volume throughput at the lowest per-token rate the provider offers. **Gemini 2.5 Flash-Lite** — $0.075 input / $0.300 output per 1M, 1M context, quality 55, tier **fast**. fast tier is built for high-volume throughput at the lowest per-token rate the provider offers. All five attributes (input, output, context, quality, tier) live in the Compare Prices grid, which makes the cross-tier deltas obvious. Stepping from Gemini 2.5 Pro (Q78, $1.25) down to Gemini 2.5 Flash-Lite (Q55, $0.075) saves 94% on input at a cost of 23 quality points — the right tradeoff if your workload tolerates the quality drop.
Where Each Google Model Earns Its Place
**Gemini 2.5 Pro**: best for production routing default — chatbots, RAG answer synthesis, structured output, anything that ships to real users at scale. **Gemini 2.5 Flash**: best for high-volume classification, lightweight summarization, embeddings-adjacent tasks, prefilters and triage stages, draft generation. **Gemini 2.5 Flash-Lite**: best for high-volume classification, lightweight summarization, embeddings-adjacent tasks, prefilters and triage stages, draft generation. This isn't marketing copy — it's how the tier classification on TokenRate's filter panel actually slots them. If you've already filtered to a tier, the Compare Prices grid is the next step: check the relevant Google models alongside their cross-provider peers (e.g., Gemini 2.5 Flash-Lite next to Gemini Flash, or Gemini 2.5 Pro next to GPT-5 and Grok 4) to confirm you're not paying a provider premium.
Cost Multipliers When Stepping Up a Tier
Within Google's lineup, the step-up multipliers are stark. Input: Gemini 2.5 Flash-Lite → Gemini 2.5 Pro is 16.7×. Output: 33.3×. Quality: +23 points. The question to ask: is +23 quality points worth 16.7× the per-token cost? For agentic or accuracy-critical workloads, yes — quality wins are non-linear in user value. For high-volume classification or templated content, no — the cheaper model clears the bar. Use the API cost estimator to put a dollar figure on the step-up at your workload volume.
Compare-Prices Across Providers, Not Just Within
The most common mistake when picking within Google's lineup is forgetting that cross-provider competitors may dominate the chosen tier. Once you've picked your Google candidates, add 1-2 competitors from a different provider dropdown — for the balanced tier, comparing Google's mid-model against Gemini 2.5 Pro and Claude Sonnet 4.7 is usually instructive. The Compare Prices grid was designed for exactly this multi-provider workflow. Both the price denominator (OpenRouter) and the quality numerator (Arena AI + Artificial Analysis) refresh hourly. So the comparison you screenshot Monday morning is still trustworthy at standup Tuesday morning — but you should re-run it before a quarterly model-routing review. Try the comparison yourself at /tools/compare-prices — it's the fastest way to stack model cost, context, and quality in a single grid.