bit.fanBit.FanWeb3 News · Betting Reviews · Market Insights

AI programming

Specific Labs
2026-09-11 09:30:11

Real-SWE tests coding agents on private enterprise code, with the top pass rate reaching only 38.8%

Specific Labs, an AI data company backed by Y Combinator, has launched Real-SWE, a programming benchmark designed to test coding agents against real private codebases from operating companies. The tasks come directly from production environments and include tax calculation, billing migration, API billing, and customer data migration. Because neither the code nor the answers are publicly available online, agents have to work out each company’s business rules and code structure on their own. In the benchmark results shared in the report, Fable 5.1 ranked first with a 38.8% pass rate. GPT-6 Astra followed at 33.8%, while Gemini 3.8 Flash posted 31.2%. GLM 5.3 came in fourth at 28.8%, outperforming Grok 4.6, Kimi K3, and GPT-5.6 Sol. Among the 10 tasks currently disclosed, six had overall pass rates below 15%, and one task saw every model fail. According to Xiaopu Peng, a researcher at Zhipu, the team has been training long-horizon tasks since GLM-5.1. He said Real-SWE is closer than public SWE leaderboards to measuring that type of capability, which may help explain GLM 5.3’s stronger showing in this benchmark.

1280
Real-SWE tests coding agents on private enterprise code, with the top pass rate reaching only 38.8%
Cognition AI
2026-09-02 02:01:10

Cognition AI Nears New Funding Round at About $47B Valuation

Cognition AI, a startup that applies artificial intelligence to software programming, is close to completing a new funding round that would lift its valuation to approximately $47 billion, according to people familiar with the matter. The latest round is expected to raise about $1 billion, the people said. Because investor demand has far exceeded the company's early expectations, the final size of the round could surpass $1 billion, they added. Cognition has already received nearly $10 billion in investment interest, the sources noted, though they did not identify the investors. Negotiations are still under way, and the specific terms could still change before the deal is finalized. BlockBeats reported the development on September 2. The company's previous valuation was not disclosed in the report, and no further details about the round's structure were available.

980
Cognition AI Nears New Funding Round at About $47B Valuation
Claude Code
2026-06-30 19:01:42

Claude Code Next Version Enables Background Sub-Agents by Default: Auto-Refactoring, Testing & PRs While You Chat

Claude Code announces that its next version will enable background sub-agents by default, allowing users to auto-execute code refactoring, testing, and PR submissions while chatting. This marks a shift from a conversational tool to a multi-task workflow engine. Spotify and other enterprises have validated its efficiency on tens of millions of lines of code, shifting engineer focus toward product decisions and requirements definition.

770
Claude Code Next Version Enables Background Sub-Agents by Default: Auto-Refactoring, Testing & PRs While You Chat