Vijay Bharadwaj, if they don't bring anything more, I agree they might not be priority 0, I'd like perplexity though.

Vijay Bharadwaj Thanks for the response! My responses are below for each bullet. Thanks!

• I understand this, and that just might be due to OpenAI and Anthropic getting the most attention and a lot of users stick with models they are most familiar with hearing about. However, like me, I think there are probably a lot your users that keep up new model releases and checkpoints that want greater access to a variety of models to test and use for their specific use cases.

• I understand this in general, especially with regard to “ability.” However, I think cost may be a different story. For example, in Mistral Medium 3’s release they claim, “Mistral Medium 3 delivers frontier performance while being an order of magnitude less expensive. For instance, the model performs at or above 90% of Claude Sonnet 3.7 on benchmarks across the board at a significantly lower cost ($0.4 input / $2 output per M token).” Claude Sonnet 3.7 is an incredible model that I have used a lot. At 40 credits, it’s not a model that I can use consistently though. If there’s any truth behind Mistral’s claim with Medium 3 and maybe it costs 5 credits instead with similar capability, this is a big deal. In the release they back that up with, “Mistral Medium 3 introduces a new class of models that balances… performance 8X lower cost.” Again, this is a big deal and would be a great feature for your customers to have access to options just like this.

• I understand. I do appreciate this challenge that you have. However, as your customers have raised several times across the forum here, we’d like some sort of ability to “favorite” our preferred models, and those would appear at the top, or something like that. I’m not a UX expert, but there has got to be a way to offer additional models and capabilities while still maintaining clean UI/UX.

Thank you so much for your transparency and feedback here. Please let me know if you have any questions or additional thoughts here. Thank you!

Vijay Bharadwaj in in my experience each model has its own unique ways and modes, and I think that it’s an injustice to limit our ability to explore. If I just wanted a subscription for the corporate Giants, I would go to any other multi AI platform that only has open AI anthropic deep seek and grok. for me what made Merlin unique was that I could explore other options in a environment that didn’t feel so corporate.

k nabu, Brent Farrington Got it. I talked to the team -- maybe we can't take it on a P0 basis, since we're in process of fixing a lot of agentic chat related bugs and issues, but we'll pick it up when we have bandwidth!

Arnaud: Perplexity's Sonar is a pretty weak offering with limited context window and limited support for function calling, which makes it incompatible with our current chat capabilities.

I'd suggest you to use Gemini 2.5 Flash (thinking) with Live Search on, on Merlin Chat. In my experience, I've used both Merlin and Perplexity side by side for testing, and Merlin literally beats Pro Search when it comes to complex searches, scraping like requests, etc.

But you be the judge, please! Try it out and lemme know how it goes!