[AINews] not much happened today
Quick Answer
DeepSeek's V4-Flash 0731 API launch marks a significant leap in performance, achieving a Terminal-Bench score of 82.7 without architectural changes, while offering a competitive cost of $0.28 per million output tokens.
Quick Take
The model's open weights are now available, enhancing integration into existing coding stacks and positioning it firmly on the Pareto frontier against proprietary systems.
Key Points
- DeepSeek V4-Flash 0731 achieves a score of 82.7, up 25.8 from April.
- Cost per million output tokens is set at $0.28, competitive with proprietary systems.
- Open weights released under MIT, enabling local deployment with specific RAM requirements.
- Model shows significant gains in agentic performance, including a 12% drop in output-token usage.
- Developers rapidly integrated DeepSeek into existing coding stacks, enhancing accessibility.
DeepSignal Analysis
What happened
DeepSeek launched its V4-Flash 0731 API, achieving a Terminal-Bench score of 82.7, a significant increase from its previous score of 56.9. The model's open weights were released under the MIT license, allowing for integration into existing coding stacks. The API offers competitive pricing at $0.28 per million output tokens, with a notable cache-hit discount.
Key evidence
- DeepSeek's V4-Flash 0731 API achieved a Terminal-Bench score of 82.7, up from 56.9 in the April preview, indicating a significant performance leap.
- The model's open weights were released on Hugging Face under the MIT license, facilitating integration into various coding environments.
- The API pricing is set at $0.28 per million output tokens, with an aggressive cache-hit discount reducing costs to $0.0028 per million cached tokens.
Why it matters
The release of DeepSeek's V4-Flash 0731 API marks a notable return to relevance for the company after a period of obscurity. The performance improvements and open weights position it competitively against proprietary models like GPT-5.6, potentially influencing market dynamics. The integration capabilities and cost-effectiveness may encourage developers to adopt DeepSeek's technology, impacting the broader AI landscape.
Source Excerpt
apart from DeepSeek V4-Flash 0731, a quiet day.
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from Latent Space
See more →
Ontologies Are So Back: Why AI Agents Are Reviving the Semantic Web
Frank Coyle at AIEWF 2026 emphasized the revival of ontologies as essential 'logical guardrails' for effective AI agents, integrating them with for better reasoning. Neo4j's Emil Eifrem highlighted three ontology types to enhance agent scalability, while Kingsley Idehen discussed the challenges and benefits of maintaining ontologies in AI systems.
![[AINews] SpaceXAI launches Grok 4.5, first Opus-class model post Cursor acquisition](https://substackcdn.com/image/fetch/$s_!8D6O!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fpbs.substack.com%2Fmedia%2FHMuQw2BXUAAJaQd.png)
