Cette section n'est pas encore traduite. Le contenu ci-dessous est affiché en anglais, tel qu'il a été publié. Lire la version anglaise.
Anthropic · Text model · Limited
Compare with similar models Open the Model Finder
Figures checked 1 Nov 2026. A dash means not published on the sources we checked, so it stays blank rather than estimated.
Anthropic names these settings effort. Price per token is the same at every level. Higher effort uses more tokens, so a task costs more. platform.claude.com (opens in a new tab)
| Level | When to use it, per Anthropic | AA index | Tokens/s |
|---|---|---|---|
| low | Most efficient. Simpler tasks that need the best speed and lowest cost, such as subagents platform.claude.com (opens in a new tab) | – | – |
| medium | Balanced. Agentic tasks that need a balance of speed, cost and performance platform.claude.com (opens in a new tab) | – | – |
| high default | Complex reasoning, difficult coding problems and agentic tasks platform.claude.com (opens in a new tab) | – | – |
| xhigh | Long-running agentic and coding tasks (over 30 minutes) with token budgets in the millions platform.claude.com (opens in a new tab) | – | – |
| max | Tasks needing the deepest possible reasoning and most thorough analysis, with no limit on token spend platform.claude.com (opens in a new tab) | – | – |
Our rule of thumb: Simple or speed-critical work: low. Standard work: the lab's default. Complex work: high. Use xhigh or max only when your own tests show a gain, as OpenAI and Anthropic both advise. developers.openai.com (opens in a new tab)platform.claude.com (opens in a new tab)
US$ per 1M tokens, standard tier, short prompts. Output left blank where not published.
| Model | Input US$/1M | Output US$/1M |
|---|---|---|
| Claude Mythos 5.1 | 10 | 50 |
| GPT-5.6 Cyber | 12.5 | 75 |
| GPT-5.5 | 5 | 30 |
| GPT-5.6 Sol | 4 | 20 |
| Kimi K3.1 Pro Highspeed | 4 | 20 |
| Lab | Anthropic |
|---|---|
| Type | Text |
| Status | Limited platform.claude.com (opens in a new tab)Anthropic limits Mythos models to approved organisations |
| Released | – |
| API access | Limited docs.anthropic.com (opens in a new tab) |
| API model ID | – |
| API rate limits | – |
| Consumer app plans | – |
| Price at each level | Same per-token price at every level. platform.claude.com (opens in a new tab) |
| Input per 1M tokens | US$10 docs.anthropic.com (opens in a new tab) |
| Output per 1M tokens | US$50 docs.anthropic.com (opens in a new tab) |
| Cached input | US$0.25 docs.anthropic.com (opens in a new tab) |
| Batch in / out | US$5 / US$25 docs.anthropic.com (opens in a new tab) |
| Other pricing | – |
| Context window | – |
| Max output | – |
| Input types | text, image, pdf platform.claude.com (opens in a new tab) |
| Output types | text |
| Weights | Closed docs.anthropic.com (opens in a new tab) |
| Reasoning | Yes platform.claude.com (opens in a new tab) |
| Tool use | Yes docs.anthropic.com (opens in a new tab) |
| Artificial Analysis index | – |
| Output speed | – |
| Time to first token | – |
| Reasoning levels | low, medium, high (default), xhigh, max platform.claude.com (opens in a new tab) |
| Lab benchmark claims | – |
| Limitations |
|
Anthropic: MCP connector: remote MCP servers in the Messages API platform.claude.com (opens in a new tab). These official remote MCP servers fit that rule. We have not tested each server with this model. A server that publishes a client list is left out when this lab is not on it.
Same type, closest on quality index and blended price. A similarity, not a ranking.
All 109 AI models · AI agents · How we check sources
Anthropic 
Fortune reports $542–567bn of compute commitments and a plan, "Anthropic Direct", to lease dedicated data centres whose partners secure land, power and buildings.
Source: Fortune (opens in a new tab)
Anthropic
A research-preview app where Claude works as a proactive agent and checks with you before acting on your behalf.
Source: Claude on X (opens in a new tab)
Anthropic
Anthropic's fastest model costs $0.10/$0.50 per 1M tokens for prompts up to 100k. Anthropic says it costs around 75% less to run on average.
Source: Anthropic (opens in a new tab)
