Observed Signal · Oct 20, 2026 · Product Launch · Source: Trending Topics (DACH/CEE Innovation & Tech) · Impact: 4/5 · Sentiment: Positive
eustella Launches Version 1.5, Boosts Open-Weight AI to Top
Vienna-based startup eustella has released version 1.5 of its AI platform, featuring improved open-weight models through proprietary post-training. The new model, based on Qwen 3.8 27B, outperforms the base model by an average of 5 points across 14 benchmarks, and rivals leading US models in coding and long-document comprehension. Notably, eustella claims a 16-fold reduction in political bias compared to Chinese base models, reducing responses reflecting Chinese government narratives from 30.4% to 1.9%. The company also introduces team and enterprise plans at €20 and €80 per user per month, respectively, offering a dedicated retrieval system, organization-wide search, and central administration. eustella now serves over 22,000 users and has granted early API access to 25 partners. The development is supported by the AI Factory Austria, part of the EuroHPC Joint Undertaking, underscoring European AI sovereignty efforts.
This is a significant product launch from a European AI startup, showcasing competitive performance against US and Chinese models while addressing critical concerns like bias and sovereignty. It highlights the growing European AI ecosystem and could influence the adoption of open-weight models in enterprise settings.
Track Mistral AI Signals & Market Shifts in Real-Time
Polaris7 autonomous intelligence agents track regulatory filings, primary sources, executive changes, and deal flow 24/7. Create your free Explorer workspace to monitor these entities.
Key Takeaways & Evidence Grounding
- eustella released version 1.5 with post-trained open-weight AI models.
- The v1.5 model, based on Qwen 3.8 27B, outperforms the base by 5 points average across 14 benchmarks.
- Political bias reduced 16-fold: responses mirroring Chinese government narratives dropped from 30.4% to 1.9%.
- eustella v1.5 achieves 89.1% on LiveCodeBench, ahead of GPT-5.6 Sol (82.6%).
- New team and enterprise plans priced at €20 and €80 per user/month.
- eustella now serves over 22,000 users and has 25 API partners.
- eustella is supported by the AI Factory Austria AI:AT, part of EuroHPC JU.
Connected Companies & Entities
4 Entities mapped“eustella v1.5 outperforms Mistral Medium 3.5 on all seven comparable benchmarks.”
“eustella presented its platform at the hub with guests from Mistral AI, Bitpanda, and IONOS Cloud.”
“eustella presented its platform at the hub with guests from Mistral AI, Bitpanda, and IONOS Cloud.”
Ontology Mapping & Concepts
Related Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
Ecosia Drops Mistral for Chinese Open-Source AI
European search engine Ecosia has switched its AI supplier from French startup Mistral to open models, including Chinese ones like Alibaba's Qwen, Z.ai's GLM, and Moonshot AI's Kimi. CEO Christian Kroll cited disappointment with Mistral's model quality, which he says lags about a year behind competitors, and frequent server overloads. Ecosia now uses models via German platform Melious, which runs open AI models on EU servers, halving AI service costs while improving performance. The switch comes as Mistral released Large 4, its most powerful model yet, trained in Europe with open weights due in October. Despite scoring 38.4 on the Intelligence Index, it ranks eighth among open models, behind seven Chinese models. The case highlights the European AI sovereignty dilemma: top-tier open models are predominantly Chinese, even when run on European servers. Mistral itself hosts Chinese models like GLM on its neocloud platform, while its CEO Arthur Mensch defends Large 4's capabilities in areas like cyber defense.
Ecosia Switches from Mistral to Chinese AI Models
Mistral AI, the Paris-based startup, released Mistral Large 4, a trillion-parameter open-weight model, scoring 38 points on Artificial Analysis and ranking 8th among open models, trailing Chinese counterparts like Xiaomi's MiMo-V2.6-Pro. Berlin-based search engine Ecosia switched from Mistral to open-weight models (Qwen, GLM, Kimi), sourced via German platform Melious, citing cost, energy efficiency, and better performance. Ecosia CEO Christian Kroll emphasized the switch is not a preference for Chinese technology but was driven by Mistral's quality lag and sovereignty concerns. Ecosia can switch models for sensitive topics to avoid political biases. The episode also covers Mistral's 'Neocloud' strategy with its app 'Vibe' and 3,800 Blackwell chips, and notes that Western open-weight models have closed the gap to closed models to 4-6 months.
Alibaba Releases Qwen3.5; Cloud Agents Overrun Local Models
Alibaba’s Qwen team open-released Qwen3.5 in four sizes (0.8B, 2B, 4B, 9B) and published full model weights on Hugging Face; the 9B variant posts benchmark results approaching much larger systems and is optimized to run on laptops and high-end phones. The newsletter argues that while efficient, open small models face a shifting competitive landscape: cloud-hosted, agentic systems (examples include OpenClaw-based agents and orchestrators running frontier models like Opus 4.6) deliver long-horizon loops, tool calls, structured memory and distilled outcomes, compounding capability beyond single-model inference. Other items covered: the U.S. Supreme Court refused to hear Stephen Thaler’s DABUS copyright appeal (leaving human-authorship requirements intact), Anthropic rolled out a limited voice mode for Claude Code, and multiple startups and product previews (Stripe billing preview, various AI tool launches) were noted.
Receive all future signals for Mistral AI & competitors in your briefing
Never miss material M&A deals, leadership changes, account wins, or product rollouts. Polaris7 monitors 20,000+ primary sources 24/7, delivering verified telemetry directly into your executive briefing.