--- title: MiMo-V2.6-Pro tops open weights models on Artificial... description: MiMo-V2.6-Pro has claimed the top spot among open weights models on the Artificial Analysis Intelligence Index with a score of 46. It keeps the same... canonical: https://daily.dev/posts/mimo-v2-6-pro-tops-open-weights-models-on-artificial-analysis-intelligence-index-oyevo1pee twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: MiMo-V2.6-Pro tops open weights models on Artificial Analysis Intelligence Index | daily.dev og:description: MiMo-V2.6-Pro has claimed the top spot among open weights models on the Artificial Analysis Intelligence Index with a score of 46. It keeps the same... og:url: https://daily.dev/posts/mimo-v2-6-pro-tops-open-weights-models-on-artificial-analysis-intelligence-index-oyevo1pee og:image: https://api.daily.dev/og/posts/oYEVo1PeE.png og:image:alt: MiMo-V2.6-Pro tops open weights models on Artificial Analysis Intelligence Index og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # MiMo-V2.6-Pro tops open weights models on Artificial Analysis Intelligence Index **[Collections](https://daily.dev/sources/collections)** · 3 min read · 4 upvotes · 0 comments ## Summary MiMo-V2.6-Pro has claimed the top spot among open weights models on the Artificial Analysis Intelligence Index with a score of 46. It keeps the same architecture as V2.5 but adds a 1M token context window and an MTP drafter, landing just behind GPT-4.5 Sol while costing $0.13 per million input tokens. ## Content Xiaomi has released MiMo-V2.6-Pro and MiMo-V2.6-Flash, two open-weight multimodal models that have quickly drawn attention both for their benchmark performance and for how openly the company shared the training process. ## Benchmark performance Artificial Analysis ranks MiMo-V2.6-Pro first among 114 tracked open-weight models on its Intelligence Index, scoring 46. It sits just below closed frontier models like Claude Opus 5.5 and GPT-6 Astra on several benchmarks, with Terminal Bench 4.0 being a notable gap. At $0.13 per task, it's also the cheapest model Artificial Analysis currently tracks. Flash, priced at roughly a third of Pro's cost, is the more practical production option for most use cases. ## Architecture Both models use Mixture-of-Experts. Pro is 1.02 trillion parameters total with 42B active; Flash is 309B total with 15B active. The architecture is the same class as V2.5: hybrid sliding-window plus global attention, with a 128-token sliding window size, native FP8 weights, 1M context, and built-in DFlash speculative decoding that drafts 7 tokens per step. Text, image, video, and audio are all handled in a single checkpoint. vLLM shipped day-0 support for both sizes. The architecture is deliberately simple - Grouped Query Attention with Sliding Window Attention - and that's worth noting. As researcher Sebastian Raschka pointed out, MiMo-V2.6 reaching the top of open-weight benchmarks with a classic attention design underlines something the field keeps relearning: most of the real progress comes from data and post-training improvements, not novel attention variants. Fancy attention mechanisms are mostly efficiency tweaks. ## What actually drove the gains The technical report is detailed enough to be worth reading carefully. Three things stand out: 1. **More agent tasks, more harnesses.** Training expanded across different task environments. DeepSWE pass@1 on held-out harnesses improved from roughly 50% to 66%. 2. **Better reward signals.** The team replaced a simple correctness verifier with an agentic grader that examines full execution traces rather than just final outputs. 3. **Large RL batches.** Each update used 1,568 prompts × 16 rollouts = 25,088 trajectories, with 2.7–3.7 billion training tokens per update. The full RL run took under six days at a combined cost of roughly $2.62–3.5 million for both variants. ## Openness beyond the weights This is where Xiaomi went further than most. The release includes: - Weights under the MIT license - Over 7,000 graded RL task environments - The full end-to-end RL training framework - Hyperparameters and cost data - A public livestream of the RL training runs for five days Researchers including Hugging Face's Thomas Wolf and Prime Intellect's Elie Bakouch specifically called out the RL environment release as valuable, since training is increasingly shifting toward reinforcement learning with verifiable rewards and those environments are hard to build. The contrast with less permissive releases from Meta, Moonshot, and Alibaba was noted by several observers. Separately, Xiaomi is named in an Anthropic report alleging misuse of Claude's output - worth knowing, though it doesn't change what's actually in the release. For anyone wanting to run full-parameter RL on the 310B Flash model, Peano AI has already demonstrated this on 1,000+ TPUs using vLLM for rollouts, with all 310B parameters moving across the ICI fabric in under 2 seconds. ## Questions this post answers ### What score did MiMo-V2.6-Pro get on the Artificial Analysis Intelligence Index? MiMo-V2.6-Pro scored 46 on the Artificial Analysis Intelligence Index, making it the top-ranked open weights model on that benchmark. It sits just behind GPT-4.5 Sol in overall ranking while being priced at $0.13 per million input tokens, a notable cost-to-performance position for an open weights release. _Developers comparing open weights model pricing and benchmarks can track releases like this one on daily.dev._ ### What changed in MiMo-V2.6-Pro compared to MiMo-V2.5? MiMo-V2.6-Pro keeps the same underlying architecture as V2.5 but adds a 1M token context window and an MTP (multi-token prediction) drafter. These additions helped push it to the top of the open weights ranking on the Artificial Analysis Intelligence Index with a score of 46. _Teams evaluating context window upgrades across model releases can follow updates like this on daily.dev._ ## Community take How the wider developer community reacted, aggregated from 1 discussion and 9 comments across x (as of 2026-09-23). **TL;DR:** Reaction focuses less on the leaderboard score itself and more on the open release of RL environments and training framework, which several see as the more durable contribution than the model weights. **Sentiment:** 55% positive · 40% mixed · 5% skeptical **The case for** - The 7,000+ task environments and RL framework are seen as harder-to-get infrastructure than the model weights themselves. - Open weights plus eval infrastructure lower the barrier for agent teams to build and test without in-house tooling. - The cost-per-task economics, if they hold up outside benchmarks, could make heavy agentic workflows more practical. **The pushback** - Some caution that a leaderboard placement is just a snapshot and want third-party reproduction on messier, real-world agentic tasks before trusting the claimed cost/quality edge. **By community** - x (positive): Replies focus on the open RL environments and training framework as the standout release, with some calling for independent verification of the cost/quality claims on real-world tasks. **Open questions** - Can outside teams reproduce the claimed cost/quality frontier using the released task suite outside Xiaomi's own serving stack? - How does the Flash model perform on latency in real-time interactive use? **Highlights** > @omarsar0 been running 2.5 in my agent lab so I was waiting on this. open weights are nice but the 7k envs are the real story, that's the part teams usually keep locked up. that's what lets the rest of us actually tinker. > — [CalebKotz63219 on x](https://x.com/CalebKotz63219/status/2102405522238398625) > @omarsar0 One more signal to watch: if third-party reruns on messy agentic workloads — not the Index tasks — reproduce that cost-per-task edge, then this is a real open frontier. A leaderboard slot is a snapshot; durable cost-per-task is the trend. > — [Bk1man on x](https://x.com/Bk1man/status/2102432159935951094) > @omarsar0 The 7K task environments and end-to-end RL framework might have more lasting impact than the model weights. Every agent team now gets production-grade eval infrastructure without building it themselves, which has been the blocker for most. > — [jatingargiitk on x](https://x.com/jatingargiitk/status/2102408316119089615) > @omarsar0 The open-source package is the story: weights, report, RL environments, and composable harnesses let others inspect the training loop. For builders, the decisive test is reproducibility: can the same task suite reproduce the cost/quality frontier outside Xiaomi’s serving stack? > — [liuzhao\_666 on x](https://x.com/liuzhao_666/status/2102405762350014862) **Source threads** - [x](https://x.com/omarsar0/status/2102404483800052021) · 0 points · 9 comments --- Tags: [#ai](https://daily.dev/tags/ai), [#llm](https://daily.dev/tags/llm) [View this post on daily.dev](https://daily.dev/posts/mimo-v2-6-pro-tops-open-weights-models-on-artificial-analysis-intelligence-index-oyevo1pee) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"MiMo-V2.6-Pro tops open weights models on Artificial Analysis Intelligence Index","url":"https://daily.dev/posts/mimo-v2-6-pro-tops-open-weights-models-on-artificial-analysis-intelligence-index-oyevo1pee","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/mimo-v2-6-pro-tops-open-weights-models-on-artificial-analysis-intelligence-index-oyevo1pee"},"datePublished":"2026-09-21T21:03:03.158Z","dateModified":"2026-09-23T20:06:15.788Z","description":"MiMo-V2.6-Pro has claimed the top spot among open weights models on the Artificial Analysis Intelligence Index with a score of 46. It keeps the same...","isAccessibleForFree":true,"articleSection":"Collections","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Collections","logo":"https://media.daily.dev/image/upload/s--fk_6ycEi--/f_auto,q_auto/v1780996001/logos/collections?_a=BAMAMiWQ0","url":"https://daily.dev/sources/collections"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/mimo-v2-6-pro-tops-open-weights-models-on-artificial-analysis-intelligence-index-oyevo1pee","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":4},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai,llm","timeRequired":"PT3M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Collections","item":"https://daily.dev/sources/collections"},{"@type":"ListItem","position":3,"name":"MiMo-V2.6-Pro tops open weights models on Artificial Analysis Intelligence Index"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/mimo-v2-6-pro-tops-open-weights-models-on-artificial-analysis-intelligence-index-oyevo1pee#faq","mainEntity":[{"@type":"Question","name":"What score did MiMo-V2.6-Pro get on the Artificial Analysis Intelligence Index?","acceptedAnswer":{"@type":"Answer","text":"MiMo-V2.6-Pro scored 46 on the Artificial Analysis Intelligence Index, making it the top-ranked open weights model on that benchmark. It sits just behind GPT-4.5 Sol in overall ranking while being priced at $0.13 per million input tokens, a notable cost-to-performance position for an open weights release. Developers comparing open weights model pricing and benchmarks can track releases like this one on daily.dev."}},{"@type":"Question","name":"What changed in MiMo-V2.6-Pro compared to MiMo-V2.5?","acceptedAnswer":{"@type":"Answer","text":"MiMo-V2.6-Pro keeps the same underlying architecture as V2.5 but adds a 1M token context window and an MTP (multi-token prediction) drafter. These additions helped push it to the top of the open weights ranking on the Artificial Analysis Intelligence Index with a score of 46. Teams evaluating context window upgrades across model releases can follow updates like this on daily.dev."}}]} ```