Claude and more
Anthropic
from $0.20 / 1M
20 models · typical $7.00 · top 95.6% GPQA, max reasoning
Developer head to head · prices updated Oct 31, 2026
Anthropic vs Google
Claude vs Gemini: every model's API price, compared
Google starts cheaper ($0.05 vs $0.20 per 1M tokens); Google's typical model costs less ($1.50 vs $7.00); Google has the more accurate top model (96.4% vs 95.6% GPQA).
Gemini and more
from $0.05 / 1M
27 models · typical $1.50 · top 96.4% GPQA, high reasoning
- Cheaper entry modelGoogle
- More accurate top modelGoogle
- Cheaper 85%+ GPQA modelAnthropic
- More modelsGoogle
The matchups that matter
| Anthropic | ||
|---|---|---|
| Cheapest model | Claude Haiku 5.5 $0.20 via Anthropic · 89.6% GPQA, max reasoning | Gemma 3 4B IT $0.05 via Amazon Bedrock · 23.2% GPQA |
| Most accurate (GPQA) | Claude Sonnet 5.5 $4.00 via Anthropic · 95.6% GPQA, max reasoning | Gemini 4 Argon $4.00 via Google Gemini API · 96.4% GPQA, high reasoning |
| Cheapest with 85%+ GPQA | Claude Haiku 5.5 $0.20 via Anthropic · 89.6% GPQA, max reasoning | Gemini 3 Flash Preview $1.13 via Google Gemini API · 89.4% GPQA, high reasoning |
| Typical model price | $7.00 | $1.50 |
| Models we track | 20 | 27 |
Flagships: Claude Sonnet 5.5 vs Gemini 4 ArgonCheapest: Claude Haiku 5.5 vs Gemma 3 4B ITClaude Opus 5.5 vs Gemini 4 ArgonGemini 3.7 Flash vs Claude Haiku 5.5
Blended list price per 1M tokens (3 input : 1 output) at each model's cheapest provider; only models at least two providers sell. GPQA Diamond by Epoch AI (CC BY 4.0), best measured reasoning setting.
Every model, cheapest first
Anthropic
- Claude Haiku 5.589.6% GPQA, max reasoning$0.20 via Anthropic
- claude-3-5-haiku$1.60 via Amazon Bedrock
- Claude Haiku 4.5$2.00 via Anthropic
- Claude Sonnet 590.5% GPQA, xhigh reasoning$4.00 via Anthropic
- Claude Sonnet 5.595.6% GPQA, max reasoning$4.00 via Anthropic
- Claude Sonnet 4.5$6.00 via Anthropic
- Claude Sonnet 4.683.3% GPQA, high reasoning$6.00 via Anthropic
- claude-3-5-sonnet$6.00 via Amazon Bedrock
- claude-3-7-sonnet$6.00 via Replicate
- claude-4-sonnet$6.00 via Replicate
- Claude Opus 5.590.6% GPQA, max reasoning$8.00 via Anthropic
- Claude Opus 4.5$10.00 via Anthropic
- Gemma 3 4B IT23.2% GPQA$0.05 via Amazon Bedrock
- Gemma 3 12B IT39.5% GPQA$0.0625 via Novita AI
- Gemma-3-27b-it47.7% GPQA$0.10 via DeepInfra
- Gemini 2.5 Flash-Lite$0.1313 via Oracle Cloud Generative AI
- Gemma 4 26B A4B IT73.2% GPQA, minimal reasoning$0.1375 via DeepInfra
- Gemma 4 31B IT75.8% GPQA, minimal reasoning$0.16 via W&B Inference (CoreWeave)
- Gemma 4 31B IT$0.205 via Amazon Bedrock
- Gemini 2.5 Flash$0.2625 via Oracle Cloud Generative AI
- Gemini 3.1 Flash Lite81.8% GPQA, high reasoning$0.5625 via Google Gemini API
- Gemini 3.5 Flash Lite83.3% GPQA, high reasoning$0.85 via Google Gemini API
- Gemini 3 Flash Preview89.4% GPQA, high reasoning$1.13 via Google Gemini API
- Gemini 3.6 Flash94.1% GPQA, high reasoning$1.50 via Google Gemini API
Questions people ask
Is Anthropic cheaper than Google?
No at the cheap end: Google's cheapest model, Gemma 3 4B IT, starts at $0.05 per 1M tokens, against $0.20 for Anthropic's Claude Haiku 5.5. The typical (median) model costs $7.00 at Anthropic and $1.50 at Google.
Which has the more accurate model, Anthropic or Google?
Anthropic's most accurate model we track, Claude Sonnet 5.5, scores 95.6% GPQA, max reasoning (GPQA Diamond, Epoch AI) for $4.00 per 1M tokens; Google's Gemini 4 Argon scores 96.4% GPQA, high reasoning for $4.00.
How many Anthropic and Google models can I buy through APIs?
We track 20 Anthropic and 27 Google models that at least two providers sell, with prices from each provider's pricing page.