1# Models
2
3Zed's plans offer hosted versions of major LLMs with higher rate limits than direct API access. Model availability is updated regularly.
4
5| Model | Provider | Token Type | Provider Price per 1M tokens | Zed Price per 1M tokens |
6| ---------------------- | --------- | ------------------- | ---------------------------- | ----------------------- |
7| Claude Opus 4.5 | Anthropic | Input | $5.00 | $5.50 |
8| | Anthropic | Output | $25.00 | $27.50 |
9| | Anthropic | Input - Cache Write | $6.25 | $6.875 |
10| | Anthropic | Input - Cache Read | $0.50 | $0.55 |
11| Claude Opus 4.1 | Anthropic | Input | $15.00 | $16.50 |
12| | Anthropic | Output | $75.00 | $82.50 |
13| | Anthropic | Input - Cache Write | $18.75 | $20.625 |
14| | Anthropic | Input - Cache Read | $1.50 | $1.65 |
15| Claude Sonnet 4.5 | Anthropic | Input | $3.00 | $3.30 |
16| | Anthropic | Output | $15.00 | $16.50 |
17| | Anthropic | Input - Cache Write | $3.75 | $4.125 |
18| | Anthropic | Input - Cache Read | $0.30 | $0.33 |
19| Claude Sonnet 4 | Anthropic | Input | $3.00 | $3.30 |
20| | Anthropic | Output | $15.00 | $16.50 |
21| | Anthropic | Input - Cache Write | $3.75 | $4.125 |
22| | Anthropic | Input - Cache Read | $0.30 | $0.33 |
23| Claude Sonnet 3.7 | Anthropic | Input | $3.00 | $3.30 |
24| | Anthropic | Output | $15.00 | $16.50 |
25| | Anthropic | Input - Cache Write | $3.75 | $4.125 |
26| | Anthropic | Input - Cache Read | $0.30 | $0.33 |
27| Claude Haiku 4.5 | Anthropic | Input | $1.00 | $1.10 |
28| | Anthropic | Output | $5.00 | $5.50 |
29| | Anthropic | Input - Cache Write | $1.25 | $1.375 |
30| | Anthropic | Input - Cache Read | $0.10 | $0.11 |
31| GPT-5 | OpenAI | Input | $1.25 | $1.375 |
32| | OpenAI | Output | $10.00 | $11.00 |
33| | OpenAI | Cached Input | $0.125 | $0.1375 |
34| GPT-5 mini | OpenAI | Input | $0.25 | $0.275 |
35| | OpenAI | Output | $2.00 | $2.20 |
36| | OpenAI | Cached Input | $0.025 | $0.0275 |
37| GPT-5 nano | OpenAI | Input | $0.05 | $0.055 |
38| | OpenAI | Output | $0.40 | $0.44 |
39| | OpenAI | Cached Input | $0.005 | $0.0055 |
40| Gemini 3.0 Pro | Google | Input | $2.00 | $2.20 |
41| | Google | Output | $12.00 | $13.20 |
42| Gemini 2.5 Pro | Google | Input | $1.25 | $1.375 |
43| | Google | Output | $10.00 | $11.00 |
44| Gemini 2.5 Flash | Google | Input | $0.30 | $0.33 |
45| | Google | Output | $2.50 | $2.75 |
46| Grok 4 | X.ai | Input | $3.00 | $3.30 |
47| | X.ai | Output | $15.00 | $16.5 |
48| | X.ai | Cached Input | $0.75 | $0.825 |
49| Grok 4 Fast | X.ai | Input | $0.20 | $0.22 |
50| | X.ai | Output | $0.50 | $0.55 |
51| | X.ai | Cached Input | $0.05 | $0.055 |
52| Grok 4 (Non-Reasoning) | X.ai | Input | $0.20 | $0.22 |
53| | X.ai | Output | $0.50 | $0.55 |
54| | X.ai | Cached Input | $0.05 | $0.055 |
55| Grok Code Fast 1 | X.ai | Input | $0.20 | $0.22 |
56| | X.ai | Output | $1.50 | $1.65 |
57| | X.ai | Cached Input | $0.02 | $0.022 |
58
59## Usage {#usage}
60
61Any usage of a Zed-hosted model will be billed at the Zed Price (rightmost column above). See [Plans and Usage](./plans-and-usage.md) for details on Zed's plans and limits for use of hosted models.
62
63> LLMs can enter unproductive loops that require user intervention. Monitor longer-running tasks and interrupt if needed.
64
65## Context Windows {#context-windows}
66
67A context window is the maximum span of text and code an LLM can consider at once, including both the input prompt and output generated by the model.
68
69| Model | Provider | Zed-Hosted Context Window |
70| ----------------- | --------- | ------------------------- |
71| Claude Opus 4.5 | Anthropic | 200k |
72| Claude Opus 4.1 | Anthropic | 200k |
73| Claude Sonnet 4 | Anthropic | 200k |
74| Claude Sonnet 3.7 | Anthropic | 200k |
75| Claude Haiku 4.5 | Anthropic | 200k |
76| GPT-5 | OpenAI | 400k |
77| GPT-5 mini | OpenAI | 400k |
78| GPT-5 nano | OpenAI | 400k |
79| Gemini 2.5 Pro | Google | 200k |
80| Gemini 2.5 Flash | Google | 200k |
81| Gemini 3.0 Pro | Google | 200k |
82
83> Context window limits for hosted Sonnet 4 and Gemini 2.5 Pro/Flash may increase in future releases.
84
85Each Agent thread and text thread in Zed maintains its own context window.
86The more prompts, attached files, and responses included in a session, the larger the context window grows.
87
88Start a new thread for each distinct task to keep context focused.
89
90## Tool Calls {#tool-calls}
91
92Models can use [tools](./tools.md) to interface with your code, search the web, and perform other useful functions.