Sonar Pro
Jan 2024
Sonar ProCompare inon
Benchmarks of Sonar Pro4
BenchmarkByScoreRank
Overall1
Science1
Legacy · saturated or no longer reported2
No benchmark matches
Rankingin non-thinking mode?
Strong at
Alternatives?
ModelScoreCostPer taskWhat changes
CheaperClaude Sonnet 5.5max×7.4-33%$5.46No forced tool use+5
BetterGPT-6 Lunalow×2.8-97%$0.005Function calling only with none · Chat Completions+3
OpenMiMo v2.6 Prothinking×6.1-91%$0.13Pass back reasoning_content+3
CheaperGemini 4 Argonhigh×6.9-33%*$1.99Synchronous output stops at 65,536+2CheaperGPT-6.1 Solmax×6.8-33%$0.72Remove unsupported parameters+5ModelScoreCostPer taskWhat changes
Claude Sonnet 5.5max×7.4-33%$5.46No forced tool use+5Gemini 4 Argonhigh×6.9-33%*$1.99Synchronous output stops at 65,536+2GPT-6.1 Solmax×6.8-33%$0.72Remove unsupported parameters+5
Muse Spark 1.3max×6.3-67%$1.60Only auto tool_choice is supported+5GPT-6 Solmax×6.3-33%$1.04ModelScoreCostPer taskWhat changes
GPT-6 Lunalow×2.8-97%$0.005Function calling only with none · Chat Completions+3GPT-5.6 Lunalow×2.8-93%$0.01
Mistral Small 4thinkingOpen+48%-96%$0.015MiMo v2.5thinkingOpen×2.9-97%$0.019
DeepSeek V3 0324non-thinkingOpen+28%-93%$0.021Context: 200K → 128K+1ModelScoreCostPer taskWhat changes
MiMo v2.6 ProthinkingOpen×6.1-91%$0.13Pass back reasoning_content+3DeepSeek V4.1 Flash 1017maxOpen×5.9-91%$0.26Pass back reasoning_content · Chat Completions+4
GLM 5.3maxOpen×5.9-64%$2.01No image input+4
Step 5 PreviewthinkingOpen×5.7-76%$1.03No web search+3
Kimi K3maxOpen×5.7same$2.00Sampling parameters are fixed+3ModelScoreCostPer taskWhat changes
Sonar Reasoning Prothinking+55%—$2 / $8Context: 200K → 128KSonarnon-thinkingsame—$1 / $1Context: 200K → 127KSonar Pro Searchn/a—$3 / $15pplx Decider Flash v1.2n/a—$0.01 / $0Context: not verifiedpplx Decider v1.2n/a—$0.02 / $0Context: not verifiedAll Perplexity models →Benchmarks
all categoriesChat & Knowledge
Humanity's Last Exam6.6#233
AA Intelligence Index 4.3.2≈7.6#259
GPQA Diamond57.8#235
Coding
LiveCodeBench · AA archived27.5#128
FAQ
How much does Sonar Pro cost?›
$3 per 1M input tokens and $15 per 1M output tokens.
Is Sonar Pro free to use?›
There is no free API; the cheapest paid price is $3 per 1M input tokens.
How good is Sonar Pro, and what is it best at?›
#231 of 334 in Chat & Knowledge, #127 of 152 in Coding.
Is Sonar Pro better than Qwen3 30B A3B?›
On AA Intelligence Index 4.3.2, Sonar Pro scores 7.6 vs 7.6 — about the same.
How can I use Sonar Pro?›
Through the API of Perplexity.
What is the context window of Sonar Pro?›
200K tokens, with up to 8K output tokens.
When was Sonar Pro released? Is there a newer version?›
Released 2024-01-01. It has been superseded by Sonar Reasoning Pro.
Is Sonar Pro open source?›
No, Sonar Pro is a closed model available through an API. The closest open-weights option here is MiMo v2.6 Pro.
Can Sonar Pro read images?›
It accepts text and images as input.
What is Sonar Pro?›
Sonar Pro is a closed language model from Perplexity, released Jan 2024.
What are the best alternatives to Sonar Pro?›
Claude Sonnet 5.5 (cheaper), GPT-6 Luna (better), MiMo v2.6 Pro (open weights).