Anthropic Launches Claude Opus 5 (Fast) at $10/$50 per Million Tokens, 1M Context Window
Anthropic has released Claude Opus 5 (Fast), a higher-throughput variant of Opus 5 that carries identical capabilities but runs at roughly 2x the price of the standard model. The model ships with a 1 million token context window and is available now through OpenRouter.
Anthropic Ships Speed-Optimized Opus 5 Variant
Anthropic has released Claude Opus 5 (Fast), a fast-mode variant of its Opus 5 model line, according to a listing on OpenRouter. The model is priced at $10 per 1M input tokens and $50 per 1M output tokens, and supports a 1 million token context window.
According to Anthropic's documentation, the Fast variant delivers "identical capabilities" to the standard Opus 5 model but with higher output speed, at roughly 2x the pricing of the regular version. This suggests a base Opus 5 model exists at a lower price point, though Anthropic has not published separate standard-tier pricing in the material reviewed for this article.
The listing shows a release date of July 24, 2026. OpenRouter is currently hosting the model through a single provider, meaning requests are forwarded directly without cross-provider routing decisions.
What's known about pricing and performance
OpenRouter's platform tracks "effective pricing" — the average cost customers actually pay after prompt caching is applied. For models with heavy repeated context, that discount can reportedly reach 60-80% off list price, though actual savings depend on individual caching patterns and were not broken out specifically for Opus 5 (Fast) in the source material.
No benchmark scores, parameter counts, or training cutoff date were disclosed in the available listing. Throughput, latency, and time-to-first-token (TTFT) metrics are tracked live on OpenRouter's dashboard but were not included as static figures in the source content.
Access
The model is available now via OpenRouter's API, which is OpenAI-compatible — developers can typically switch to it by swapping the base URL and model slug (anthropic/claude-opus-5-fast) in existing integrations. Uptime is monitored continuously by OpenRouter, which automatically retries requests through alternate providers if the primary provider errors out — though currently only one provider serves this model.
What this means
A "Fast" variant priced at 2x the standard rate is a familiar pattern: it lets developers pay a premium for lower latency and higher throughput on the same underlying capability set, without waiting for or requesting a distinct smaller model. This is useful for latency-sensitive applications — real-time agents, interactive coding assistants, customer-facing chat — where token generation speed matters more than cost efficiency.
The bigger open question is what "Opus 5" itself represents. Its existence implies Anthropic has already released or is preparing to release a full Opus 5 line, of which this Fast variant is one configuration. Pricing at $10/$50 per 1M tokens places it in premium territory, above most current frontier model pricing, consistent with Anthropic's historical practice of pricing Opus-tier models at the top of the market. Until Anthropic publishes benchmark data, capability claims about Opus 5 remain unverified, and the practical value of the Fast variant will depend on how much latency improvement it actually delivers in production workloads.
Related Articles
Anthropic Report: Claude Was Used to Target US Navy Ships, Build Missiles, and Track Uyghurs
Anthropic's latest threat intelligence report documents five cases where state and non-state actors used Claude for military targeting, weapons development, mass surveillance, and repression. The findings include an Iran-linked operation targeting US naval forces and a Mali-based system capable of monitoring 25 million phones.
Anthropic Threat Report: Claude Used for Missile Software, Mass Surveillance, and Systematic Theft by Chinese AI Labs
Anthropic's latest threat intelligence report covers December 2025 through August 2026, documenting Claude's misuse in espionage, weapons development, and nationwide surveillance operations. The report also details how seven Chinese AI labs ran covert networks—some routing their own customers' requests through Claude—to extract training data at industrial scale.
Sakana AI Launches Fugu Ultra v2, a Multi-Agent Orchestrator With 1M-Token Context
Sakana AI has released Fugu Ultra v2, described as a learned multi-agent orchestration system rather than a single monolithic model. It offers a 1M-token context window, configurable reasoning effort, and pricing of $5 per 1M input tokens and $30 per 1M output tokens.
Analysis: Claude 'Fable 5.1' Drops Em Dashes and Hedging Language, Answers Grow 30% Longer
A new Arena.ai analysis of tens of thousands of Text Arena outputs shows Claude 'Fable 5.1' has shifted its writing style significantly from Fable 5 — using fewer em dashes, less hedging language, and producing 30% longer responses. The codenamed models appear to be unreleased Anthropic checkpoints being tested anonymously on LMArena.
Comments
Loading...