Claude vs Gemini

Anthropic's and Google's newest models sit at opposite ends of the price range. Claude Fable 5.1, released on September 1, 2026, is Anthropic's most capable generally available model and costs $10 per million input tokens and $50 per million output tokens. Gemini 3.8 Flash, released a day later, is what Google calls its best reasoning and coding model yet, at an introductory $0.75 and $3.75 through 2026. Both have a context window of about 1M tokens. Fable 5.1 reads text, images and documents, and Gemini 3.8 Flash adds video and audio input.

Anthropic aims Fable 5.1 at long, unattended work: codebase-wide features, code review and multi-hour agent runs. It keeps adaptive thinking on at all times and routes many flagged cybersecurity and biology requests to Opus models. Google aims 3.8 Flash at long-horizon coding and agents at Flash prices, notes that it can use more tokens at higher effort, and builds it into Search, Sheets and the Gemini app. On LMArena's crowd-voted text leaderboard in late September 2026, Fable 5.1 ranked a few places above Gemini 3.8 Flash, whose rating was still preliminary. The price gap is large, so check whether Flash handles your prompts before paying for Fable.