On Artificial Analysis Intelligence Index, Mimo V2.5 gets a score of 49, which is comparable to Claude 4.5 Opus at 49.7, but completes the entire benchmark with nearly half the cost of Gemini 3.1 Flash Lite (which scores 33.5 on AA Intelligence index). Here are the cost comparisons: Claude Opus 4.5: $2,969 Gemini 3.1 Pro: $892 Gemini 3.1 Flash Lite: $94 Mimo V2.5: $49 In my experience, it seems to have better follow-through than Gemini and seems less likely to say it did completed a task it didn't actually complete. And the latency is really good using Qwen CLI (a fork of Gemini CLI designed to accommodate third party models better by the Qwen team) as it runs the agentic loops really fast. There is some talk about Mimo V2.5 Pro which is $161 and scores 53.8, but I'd say that for 1/3 the price, you get most of the intelligence already and I think Mimo V2.5 pro takes the cake when doing large agentic tasks with lots of sub agents without the need of committing to a subscription.…