# Can't use high reasoning effort with anthropic/claude-opus-4-5

**URL:** <https://community.perplexity.ai/t/cant-use-high-reasoning-effort-with-anthropic-claude-opus-4-5/3866>\
**Category:** Bug Reports\
**Tags:** sonar-api\
**Created:** [February 6, 2026, 1:29pm UTC](https://community.perplexity.ai/t/cant-use-high-reasoning-effort-with-anthropic-claude-opus-4-5/3866 "2026-02-06T13:29:15Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![sszymczy](https://avatars.discourse-cdn.com/v4/letter/s/f07891/32.png) [@sszymczy](https://community.perplexity.ai/u/sszymczy)\
**Post date:** [February 6, 2026, 1:29pm UTC](https://community.perplexity.ai/t/cant-use-high-reasoning-effort-with-anthropic-claude-opus-4-5/3866/1 "2026-02-06T13:29:15Z")

</div>

## 🐛 Describe the Bug

When calling “anthropic/claude-opus-4-5" model via agentic research API with “high” reasoning effort the API returns 400 “invalid request” error. It works without issues with reasoning effort set to “low” and “medium”.

## ✅ Expected Behavior

API request completed successfully.

## ❌ Actual Behavior

API returned 400 “invalid request” error.

## 🔄 Steps to Reproduce

```auto
$ export PERPLEXITY_API_KEY="...key value..."
$ curl https://api.perplexity.ai/v1/responses \
  -H "Authorization: Bearer $PERPLEXITY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "anthropic/claude-opus-4-5",
    "input": "calculate 2+2=",
    "reasoning": {
      "effort": "high"
    }
  }'

```

## 📌 API Request & Response (if applicable)

Request:

```auto
{
    "model": "anthropic/claude-opus-4-5",
    "input": "calculate 2+2=",
    "reasoning": {
        "effort": "high"
    }
}

```

response:

```auto
{"error":{"message":"invalid request","type":"invalid_request","code":400}}

```

## 🌍 Environment

- **API Version:** Current (no idea how to check)
- **SDK (if applicable):** N/A
- **Operating System:** Linux

## 📎 Logs or Screenshots (if applicable)

Trying all three reasoning effort values:

```auto
$ curl https://api.perplexity.ai/v1/responses -H "Authorization: Bearer $PERPLEXITY_API_KEY" -H "Content-Type: application/json" -d '{
    "model": "anthropic/claude-opus-4-5",
    "input": "calculate 2+2=",
    "reasoning": {
      "effort": "low"
    }
  }'
{"created_at":1770384358,"id":"resp_9d79fcd9-a36c-4cdf-8747-4ecd3a643c38","model":"anthropic/claude-opus-4-5","object":"response","output":[{"content":[{"text":"2 + 2 = **4**","type":"output_text"}],"id":"msg_11c3891e-87ef-4180-9ca8-978706bc8f1b","role":"assistant","status":"completed","type":"message"}],"status":"completed","usage":{"cost":{"currency":"USD","input_cost":0.0005,"output_cost":0.00145,"total_cost":0.00195},"input_tokens":99,"output_tokens":58,"total_tokens":157}}
$ curl https://api.perplexity.ai/v1/responses -H "Authorization: Bearer $PERPLEXITY_API_KEY" -H "Content-Type: application/json" -d '{
    "model": "anthropic/claude-opus-4-5",
    "input": "calculate 2+2=",
    "reasoning": {
      "effort": "medium"
    }
  }'
{"created_at":1770384371,"id":"resp_43b3499f-f2be-4e52-8ada-1a5a7a73ecea","model":"anthropic/claude-opus-4-5","object":"response","output":[{"content":[{"text":"2 + 2 = **4**","type":"output_text"}],"id":"msg_d43ef1f1-6eee-4937-91ed-80748b52b079","role":"assistant","status":"completed","type":"message"}],"status":"completed","usage":{"cost":{"currency":"USD","input_cost":0.0005,"output_cost":0.00125,"total_cost":0.00175},"input_tokens":99,"output_tokens":50,"total_tokens":149}}
$ curl https://api.perplexity.ai/v1/responses -H "Authorization: Bearer $PERPLEXITY_API_KEY" -H "Content-Type: application/json" -d '{
    "model": "anthropic/claude-opus-4-5",
    "input": "calculate 2+2=",
    "reasoning": {
      "effort": "high"
    }
  }'
{"error":{"message":"invalid request","type":"invalid_request","code":400}}

```

---

<div class="post-metadata">

**Author:** ![Fatima\_Shahbaz](https://sea1.discourse-cdn.com/flex001/user_avatar/community.perplexity.ai/fatima_shahbaz/32/3257_2.png) [@Fatima\_Shahbaz](https://community.perplexity.ai/u/Fatima_Shahbaz)\
**Post date:** [June 23, 2026, 9:43pm UTC](https://community.perplexity.ai/t/cant-use-high-reasoning-effort-with-anthropic-claude-opus-4-5/3866/2 "2026-06-23T21:43:07Z")

</div>

Il [passaggio da lordo](https://stipendionettocalcolatore.it/stipendio-medio-milano/) a netto rappresenta il processo attraverso il quale si determina l’importo effettivamente ricevuto dal lavoratore a partire dalla retribuzione lorda. Durante questo calcolo vengono sottratti i contributi previdenziali, le imposte sul reddito e le eventuali addizionali locali previste dalla normativa fiscale.