Support for deepseek-R1-0528 as an agent

Feature Requests
1

Currently, a warning appears when trying to use the model as an agent

Model deepseek-r1-0528 doesn't have good agent support yet. Start a new thread using Manual mode for better results.

At the same time, this model is inexpensive in terms of the cost of tokens. And one of the top models on the Live bench.

Token Price Comparison Table (July 2025)

Model Input Tokens (per 1M) Output Tokens (per 1M) Comments
Claude Sonnet 4 $3 $15 High quality, industry standard
GPT-3 $1.5 $2 Lowest price among GPT models, but outdated
GPT-4o mini $0.15 $0.60 One of the lowest prices among top-tier models
Gemini 2.5 Pro (<200k) $1.25 $10 Price for short prompts
Gemini 2.5 Pro (>200k) $2.50 $15 Price for long prompts
DeepSeek R1 $0.10–$0.20 $0.20–$0.30 Among the most affordable models
Qwen3 (0.6B) $0.11 $0.42 Lowest cost among all models
Model Organization Global Average Reasoning Average Coding Average Agentic Coding Average Mathematics Average Data Analysis Average Language Average IF Average
o3 Pro High OpenAI 74.72 94.67 76.78 31.67 84.75 69.40 79.88 85.87
o3 High OpenAI 74.61 94.67 76.71 36.67 85.00 67.02 76.00 86.17
Claude 4 Opus Thinking Anthropic 72.93 90.47 73.25 33.33 88.25 70.73 73.72 80.74
Claude 4 Sonnet Thinking Anthropic 72.08 95.25 73.58 30.00 85.25 69.84 70.19 80.43
o3 Medium OpenAI 71.98 91.00 77.86 28.33 80.66 68.19 73.48 84.32
o4-Mini High OpenAI 71.52 88.11 79.98 28.33 84.90 68.33 66.05 84.96
Gemini 2.5 Pro (Max Thinking) Google 70.95 94.28 73.90 20.00 84.19 71.50 75.44 77.35
DeepSeek R1 (2025-05-28) DeepSeek 70.10 91.08 71.40 26.67 85.26 71.54 64.82 79.95
Gemini 2.5 Pro Google 69.39 93.72 70.70 13.33 83.33 71.60 74.52 78.54
Claude 3.7 Sonnet Thinking Anthropic 67.43 76.17 73.19 25.00 79.00 69.11 68.27 81.25
o4-Mini Medium OpenAI 66.87 78.47 74.22 21.67 81.02 68.47 62.41 81.83

Using this model DeepSeek-R1-0528 for most tasks would enhance the user experience and reduce the cost of spending in auto mode.

I look forward to her support =)

2

Hey, thanks for the feature request. Actually, we’ve already received a feature request for this earlier, and I’ve logged it in our internal tracker so the team can take a look at it.

1 Like