Swedish Startup ‘Incredible’ Announces Agentic AI Models That Can Take 1000+ Actions at Once

Sep 9, 2025

JÖNKÖPING, Sweden – September 9, 2025 – Incredible, a Swedish AI startup, today announced Incredible Small 1.0, the first in a new class of agentic AI models that outperform GPT-5 and Claude on real-world business tasks.

The breakthrough Live-Code architecture enables AI to execute hundreds of simultaneous actions across business applications—from updating entire CRM databases to processing gigabytes of financial data—while maintaining over 99% accuracy on data-critical tasks.

"Everyone's building AI to be smarter, but businesses need AI to be more useful," explains Philip Alm, CEO and co-founder. "Claude can write brilliant code, ChatGPT excels at reasoning and creative writing, but when you need to update 1,000 CRM records or process large customer data? Current LLMs hit a wall."

AI that does real work

While existing AI models excel at conversation and analysis, they struggle with the operational tasks businesses actually need: processing large amounts of company data without errors, completing multi-step workflows reliably, and working directly with business applications.

Incredible Small 1.0 changes this by handling data differently than traditional AI models. Instead of being limited to small data samples, it can work with entire business databases and execute hundreds of actions simultaneously across different applications.

The result is AI that can update thousands of customer records, analyze complete financial datasets, or handle complex business processes from start to finish - all while maintaining the accuracy that companies require.

"If you want an LLM to write code, use Claude. If you want PhD-level intelligence, use GPT. If you want AI that takes actions in products and gets work done, use Incredible," said Emil Wagman, Head of Development at Incredible.

From Language to Action: The Journey

The VC-backed 13-person Incredible team previously developed SWE-LLM in 2024—Sweden's largest fine-tuned language model trained on a carefully curated dataset consisting of over 12,700 Swedish texts. That experiment in language understanding laid the groundwork for their current breakthrough.

"SWE-LLM taught us that understanding language is just the first step," explains Alm. "The real value comes when AI can act in our applications on that understanding. We hope this opens eyes to what's possible when you focus on execution over intelligence."

The company, winner of Sweden's Innovation Award SKAPA for Young Innovators in memory of Alfred Nobel, brings deep expertise from work with Fortune 500 companies including Toyota and Volvo.

Competitive Performance & Benchmarks

Even though no formal benchmarks exist yet for parallel execution and massive data handling capabilities, these innovations translate to strong performance on traditional benchmarks. Incredible Small 1.0 outperforms leading models on real-world tasks and matches top level performance at other public benchmarks:
– TAU-bench Retail: 74.7% (vs GPT-5's 71.3% and Claude 4.0 Sonnet's 56.5%)
– TAU-bench Airline: 60% (vs GPT-5's 40% and Claude 4.0 Sonnet's 56%)
– IFEval Prompt: 86.32%
– IFEval Instruction: 90.53%
– MMLU-Pro: 84%
– GPQA: 78%

"Despite very limited resources, we've created models that compete with world-leading AI companies where it matters most: getting real work done," said Philip Alm, co-founder and CEO of Incredible. "We spent three years solving a different problem than everyone else. While others optimize for general intelligence with billions in funding, we optimize for how much real work models can actually complete.”

The Future of Swedish AI

"We want Sweden to become a leading player in the AI market, not a follower," says Alm. "For that to happen, we need to create our own models that solve real problems. We're showing that innovation doesn't require billions in resources — just different thinking about what AI should actually do."

Availability and Pricing

Incredible Small 1.0 is available now in early research preview at $1.15 per million input tokens and $5.00 per million output tokens—a 67% cost reduction on output compared to Claude Sonnet and 50% reduction compared to GPT-5. When accounting for data processing efficiency improvements, the cost savings are even more significant.

The company plans to release Incredible Big, its flagship model, in the coming weeks with substantial improvements across all benchmarks.

About Incredible

Incredible builds the world's most capable agentic AI models for taking actions in applications. Founded in 2022 in Jönköping, Sweden, the company serves enterprises, SMBs and developers who need AI that completes real work across internal and external applications.