In a single 24-hour stretch, Anthropic, Meta, and Google all shipped new frontier models — but instead of chasing raw intelligence records, two of the three led with something else entirely: cost. Meta’s headline wasn’t a benchmark score, it was a 25% cut in token usage. Google’s pitch wasn’t smarter reasoning, it was $0.75 per million tokens. The message from Big Tech is clear: the next phase of the artificial intelligence race is about doing more work for less money — not just doing harder work.
Table of Contents
3 Frontier AI Models Launches
| Model | Company | Headline Feature |
|---|---|---|
| Claude Fable 5.1 | Anthropic | 75% cheaper cache reads, stronger long-running agent work |
| Muse Spark 1.3 | Meta | 25% fewer tokens, 20% fewer tool calls, same price as predecessor |
| Gemini 3.8 Flash | $0.75/M input tokens, matches 3.7 Flash pricing with better benchmarks | |
| GPT-5.6 (context) | OpenAI | Frontier model already live, cited across all three rivals’ comparisons |
Breaking Down Each Release
Claude Fable 5.1 keeps Anthropic’s pricing steady at $10/$50 per million tokens but slashes cache-read costs by 75%, alongside major gains in long-running, tool-using agent tasks — its Terminal-Bench-Science score more than doubled versus its predecessor
Muse Spark 1.3 is Meta’s fourth Muse Spark release in five months, and its pitch is efficiency over raw power: roughly 25% fewer tokens and 20% fewer tool calls than the previous version, at unchanged pricing. Meta’s AI chief claims it now rivals Anthropic and OpenAI on coding tasks.
Gemini 3.8 Flash holds its price exactly at $0.75 per million input tokens and $3.75 per million output tokens — identical to its predecessor — while improving benchmark scores across coding and agentic tasks, undercutting most frontier competitors by a wide margin.
Adding to the momentum, OpenAI has confirmed its next model, codenamed Astra, is coming soon, with early benchmark chatter already generating buzz across the AI community.
For more tech and AI updates as this fast-moving story develops, check out our Technology section on TechnoSports.
FAQs
Q1: Why are AI companies focusing on cost instead of raw power now?
Because most frontier models already perform well — the real competitive edge now is doing the same work cheaper and faster.
Q2: Which new AI model is the cheapest to use?
Gemini 3.8 Flash currently offers the lowest per-token pricing among the three releases.
