Open-Source AI

Today's Stories

1. 1

   ComplianceAgent: Open-source EU AI Act compliance scanner

   ComplianceAgent: Open-source EU AI Act compliance scanner

2. 2

   Show HN: Open-source phone calling infra for AI agents

   Show HN: Open-source phone calling infra for AI agents

3. 3

   If your GPU can run inference, it should be able to fine-tune too. [P]

   If your GPU can run inference, it should be able to fine-tune too. [P]

4. 4

   We'll benchmark an Open weights LLM on any GPU you choose — drop your model + hardware and we'll run it. [D]

   We'll benchmark an Open weights LLM on any GPU you choose — drop your model + hardware and we'll run it. [D]

5. 5

   Open-source tool cuts Claude, GPT token costs 59–70% by hiding text in images

   Developer Steven Chong released pxpipe, an open-source proxy that converts long text inputs—system prompts, code, chat history—into compressed PNG images before sending them to Claude Code and other AI models. The trick exploits Anthropic's pricing: text costs roughly one token per character, but images cost a fixed number of tokens regardless of content density, allowing about 3.1 characters per image token. In one Fable 5 demo, session costs fell from $42.21 to $6.06. For businesses and developers using Claude Code or GPT 5.6 at scale, token costs directly affect operational spending. This technique could reduce invoices by 59 to 70 percent on many workflows. However, the approach is lossy—exact strings like hashes can become garbled when read from images, and processing is slower because the model must run images through a vision encoder instead of reading text directly. Fable 5 hits 100 percent accuracy on math benchmarks, but Opus 4.7 and 4.8 misread about 7 percent of rendered images, and GPT 5.5 also performs worse with image context.

   pxpipe supports Claude Fable 5 and GPT 5.6 by default; Opus 4.7, 4.8, and GPT 5.5 can be enabled manually. If this technique catches on widely, AI companies may respond by raising image processing prices. The approach is not new—Deepseek built a similar OCR system that compresses text documents by up to a factor of ten while retaining 97 percent of information.

6. 6

   Mistral AI eyes €1.7 billion Series C, claims path to $1 billion（約1600億円） ARR

   French AI startup Mistral closed a €1.7 billion Series C round (about $2 billion（約3200億円）) led by ASML in September 2025. The company also disclosed annual recurring revenue above $400 million（約640億円） in February, up from $20 million（約32億円） one year prior, and stated it is on track to surpass $1 billion（約1600億円） in ARR this year. Mistral is rumored to be raising some $3.5 billion（約5600億円） at a $23.15 billion（約3.7兆円） valuation. Mistral's revenue growth and funding trajectory position it as a viable alternative to U.S. AI labs in Europe, even as it pursues a lower valuation than U.S. frontier labs. The company is following a Palantir-style model, deploying engineers to help governments and large corporations adopt and customize AI—an approach better suited to its current means. This reflects broader calls for sovereign tech that reduces reliance on the U.S.

   Mistral plans to launch an open-weight language model this summer with early access in July. The company is also building a European AI cloud infrastructure through a €4 billion investment strategy (around $4.56 billion（約7300億円）) to construct data centers in France and Sweden, and will launch Mistral Compute, a European platform powered by Nvidia processors, in 2026.