### Recommended effort levels for Claude Opus 5.5
The whole hunk
from line 13, old and new numbered
/
lines
from line 13
1313 - claude-fable-5
1414 - claude-mythos-5
1515 - claude-mythos-preview
16 - claude-opus-5-5
1617 - claude-opus-5
1718 - claude-opus-4-8
1819 - claude-opus-4-7
from line 46
4546 -H "anthropic-version: 2023-06-01" \
4647 -H "content-type: application/json" \
4748 -d '{
48 "model": "claude-opus-5",
49 "model": "claude-opus-5-5",
4950 "max_tokens": 4096,
5051 "messages": [{
5152 "role": "user",
from line 60
5960
6061 ```bash CLI
6162 ant messages create \
62 --model claude-opus-5 \
63 --model claude-opus-5-5 \
6364 --max-tokens 4096 \
6465 --output-config '{effort: medium}' \
6566 --message '{role: user, content: "Analyze the trade-offs between microservices and monolithic architectures"}' \
from line 72
7172 client = anthropic.Anthropic()
7273
7374 response = client.messages.create(
74 model="claude-opus-5",
75 model="claude-opus-5-5",
7576 max_tokens=4096,
7677 messages=[
7778 {
from line 92
9192 const client = new Anthropic();
9293
9394 const response = await client.messages.create({
94 model: "claude-opus-5",
95 model: "claude-opus-5-5",
9596 max_tokens: 4096,
9697 messages: [
9798 {
from line 116
115116
116117 var parameters = new MessageCreateParams
117118 {
118 Model = Model.ClaudeOpus5,
119 Model = Model.ClaudeOpus5_5,
119120 MaxTokens = 4096,
120121 Messages = [
121122 new() {
from line 138
137138 client := anthropic.NewClient()
138139
139140 response, err := client.Messages.New(context.TODO(), anthropic.MessageNewParams{
140 Model: anthropic.ModelClaudeOpus5,
141 Model: anthropic.ModelClaudeOpus5_5,
141142 MaxTokens: 4096,
142143 Messages: []anthropic.MessageParam{
143144 anthropic.NewUserMessage(anthropic.NewTextBlock("Analyze the trade-offs between microservices and monolithic architectures")),
from line 164
163164 AnthropicClient client = AnthropicOkHttpClient.fromEnv();
164165
165166 MessageCreateParams params = MessageCreateParams.builder()
166 .model(Model.CLAUDE_OPUS_5)
167 .model(Model.CLAUDE_OPUS_5_5)
167168 .maxTokens(4096L)
168169 .addUserMessage("Analyze the trade-offs between microservices and monolithic architectures")
169170 .outputConfig(OutputConfig.builder()
from line 187
186187 messages: [
187188 ['role' => 'user', 'content' => 'Analyze the trade-offs between microservices and monolithic architectures']
188189 ],
189 model: 'claude-opus-5',
190 model: 'claude-opus-5-5',
190191 outputConfig: ['effort' => 'medium'],
191192 );
192193
from line 202
201202 client = Anthropic::Client.new
202203
203204 message = client.messages.create(
204 model: "claude-opus-5",
205 model: "claude-opus-5-5",
205206 max_tokens: 4096,
206207 messages: [
207208 { role: "user", content: "Analyze the trade-offs between microservices and monolithic architectures" }
from line 220
219220
220221## How effort works
221222
222By default, Claude uses high effort, spending as many tokens as needed for excellent results. You can raise the effort level to `max` for the absolute highest capability, or lower it to be more conservative with token usage, optimizing for speed and cost while accepting some reduction in capability.
223Most Claude models default to high effort, spending as many tokens as needed for excellent results; Claude Opus 5.5 defaults to medium. You can raise the effort level to `max` for the absolute highest capability, or lower it to be more conservative with token usage, optimizing for speed and cost while accepting some reduction in capability.
223224
224225<Tip>
225 Setting `effort` to `"high"` produces exactly the same behavior as omitting the `effort` parameter entirely.
226 Setting `effort` to the model's default (`"medium"` on Claude Opus 5.5, `"high"` on other models) produces exactly the same behavior as omitting the `effort` parameter entirely.
226227</Tip>
227228
228229The effort parameter affects **all tokens** in the response, including:
from line 236
235236
236237### Effort levels
237238
238| Level | Description | Typical use case |
239| -------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------ |
240| `max` | Absolute maximum capability with no constraints on token spending. Available on Claude Fable 5.1, Claude Mythos 5.1, Claude Fable 5, Claude Mythos 5, Claude Mythos Preview, Claude Opus 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, Claude Sonnet 5, and Claude Sonnet 4.6. | Tasks requiring the deepest possible reasoning and most thorough analysis |
241| `xhigh` | Extended capability for long-horizon work. Available on Claude Fable 5.1, Claude Mythos 5.1, Claude Fable 5, Claude Mythos 5, Claude Opus 5, Claude Opus 4.8, Claude Opus 4.7, and Claude Sonnet 5. | Long-running agentic and coding tasks (over 30 minutes) with token budgets in the millions |
242| `high` | High capability. Equivalent to not setting the parameter. | Complex reasoning, difficult coding problems, agentic tasks |
243| `medium` | Balanced approach with moderate token savings. | Agentic tasks that require a balance of speed, cost, and performance |
244| `low` | Most efficient. Significant token savings with some capability reduction. | Simpler tasks that need the best speed and lowest costs, such as subagents |
239| Level | Description | Typical use case |
240| -------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------ |
241| `max` | Absolute maximum capability with no constraints on token spending. Available on Claude Fable 5.1, Claude Mythos 5.1, Claude Fable 5, Claude Mythos 5, Claude Mythos Preview, Claude Opus 5.5, Claude Opus 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, Claude Sonnet 5, and Claude Sonnet 4.6. | Tasks requiring the deepest possible reasoning and most thorough analysis |
242| `xhigh` | Extended capability for long-horizon work. Available on Claude Fable 5.1, Claude Mythos 5.1, Claude Fable 5, Claude Mythos 5, Claude Opus 5.5, Claude Opus 5, Claude Opus 4.8, Claude Opus 4.7, and Claude Sonnet 5. | Long-running agentic and coding tasks (over 30 minutes) with token budgets in the millions |
243| `high` | Spends as many tokens as the task needs for excellent results. The default on every model that supports effort except Claude Opus 5.5. | Complex reasoning, difficult coding problems, agentic tasks |
244| `medium` | Balanced approach with moderate token savings. The default on Claude Opus 5.5. | Agentic tasks that require a balance of speed, cost, and performance |
245| `low` | Most efficient. Significant token savings with some capability reduction. | Simpler tasks that need the best speed and lowest costs, such as subagents |
245246
246247Not every model that supports `max` supports `xhigh`.
247248
from line 264
263264
264265Reduce effort if a task completes but takes longer than necessary, or if you want a faster, more interactive working style. The same recommendations apply to Claude Mythos 5. For fuller guidance, see [Prompting Claude Fable 5](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-fable-5).
265266
267### Recommended effort levels for Claude Opus 5.5
268
269Claude Opus 5.5 supports all five effort levels, and `medium` is the default (Claude Opus 5 and earlier Opus models default to `high`, so a request that omits `effort` runs one level lower than it did on Claude Opus 5). Adaptive thinking is always on and can't be turned off, so effort is the primary control for how much the model reasons and what a request costs. Run an effort sweep on your own evals rather than carrying settings over from an earlier model, and set a large `max_tokens` at the higher levels: it's a hard limit on total output (thinking plus response text). Requests that set `thinking: {"type": "disabled"}` return a 400 error at every effort level. Claude Opus 5.5 also supports [changing effort mid-conversation](https://platform.claude.com/docs/en/build-with-claude/effort#change-effort-mid-conversation-beta) with a per-message `output_config`, which preserves the prompt cache. See [Prompting Claude Opus 5.5](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-opus-5-5).
270
266271### Recommended effort levels for Claude Opus 5
267272
268273Claude Opus 5 supports all five effort levels. **Start with `high`, the default**, and adjust based on your evals: step up to `xhigh` for demanding coding and agentic work, or to `max` when a task justifies unconstrained token spending, and use `low` and `medium` liberally as your primary control for token cost and response time wherever your evals show quality holds. If you carried effort settings over from an earlier model, run a fresh effort sweep on your evals rather than reusing them.
from line 355
350355
351356## Change effort mid-conversation
352357
353You can run later turns of a conversation at a different effort level in two ways. On Claude Fable 5.1, Claude Mythos 5.1, and Claude Opus 5, use a per-message effort change, which keeps the prompt cache. On other models, set a new top-level value on the next request, which starts the cache over.
358You can run later turns of a conversation at a different effort level in two ways. On Claude Fable 5.1, Claude Mythos 5.1, Claude Opus 5.5, and Claude Opus 5, use a per-message effort change, which keeps the prompt cache. On other models, set a new top-level value on the next request, which starts the cache over.
354359
355360### Per-message effort (beta)
356361
from line 642
637642
638643## Best practices
639644
6401. **Set effort explicitly:** The API defaults to `high`, but the right starting point depends on your model and workload.
6451. **Set effort explicitly:** The API defaults to `high` (`medium` on Claude Opus 5.5), but the right starting point depends on your model and workload.
6416462. **Use low for speed-sensitive or simple tasks:** When latency matters or tasks are straightforward, low effort can significantly reduce response times and costs.
6426473. **Test your use case:** The impact of effort levels varies by task type. Evaluate performance on your specific use cases before deploying.
6436484. **Consider dynamic effort:** Adjust effort based on task complexity. Simple queries may warrant low effort while agentic coding and complex reasoning benefit from high effort. See the next item before varying it within one conversation.