Skip to content
TheChatArchiveThe Chat Archive
Perplexitysonar-reasoning-pro-high· PerplexityHugging Face LMArena Search Arena 24K

Please generate a table listing the top 10 models from OpenRouter's monthly rankings,…

From a Chatbot Arena battleThe same person sent this prompt to two anonymous models at once. Compare it with gpt-4o-search-preview-high’s answer.
  1. User

    Please generate a table listing the top 10 models from OpenRouter's monthly rankings, along with the amount of tokens used and the estimated costs for the used tokens reported by the stats.

  2. Perplexity Assistant

    Based on the search results, here is a table of the top 10 models from OpenRouter's monthly rankings, along with their token usage and estimated costs:

    RankModelMonthly Tokens UsedEstimated CostNotes
    1Anthropic: Claude 3.7 Sonnet43.5B tokens~217.5M−217.5M - 652.5M4% of total traffic, leading in Programming category [1][4]
    2Google: Gemini 2.0 Flash34.8B tokens~8.7M−8.7M - 13M7% of total traffic [1]
    3Google: Gemini 2.5 ProNot specifiedNot specifiedGaining ground on Claude [4]
    4OpenAI: GPT-4o-miniNot specified~0.15−0.15-0.6/M tokensDominated weekly rankings [4]
    5DeepSeek: DeepSeek V3Not specified~0.27/Minput,0.27/M input, 1.10/M outputSignificantly cheaper than Claude [15]
    6Optimus AlphaNot specifiedFree (currently)Trending free model [4]
    7Meta: Llama-3-70b-instructNot specified~0.59−0.59-0.79/M tokensPopular open-source model [13]
    8OpenAI: GPT-4oNot specified~2.5−2.5-10/M tokensPremium model [13]
    9Anthropic: Claude 3.5 SonnetNot specified~3/Minput,3/M input, 15/M outputPrevious generation of Claude [15]
    10Google: Gemini 1.5 FlashNot specifiedNot specifiedStill used by some users [6]

    Note: The estimated costs are calculated based on the token pricing information available in the search results. For Claude 3.7 Sonnet, the cost estimate assumes a similar pricing structure to Claude 3.5 Sonnet (3−3-15/M tokens) and is presented as a range since the exact split between input and output tokens is not specified. The actual costs may vary significantly depending on the ratio of input to output tokens and any special pricing arrangements.

    For the Programming category specifically, Claude 3.7 Sonnet was the top performer in monthly rankings, while GPT-4o-mini dominated the weekly rankings despite Quasar Alpha's brief appearance as a free option[4]. The rankings can shift rapidly as new models are introduced, with free models like Optimus Alpha gaining momentum[4].

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.