Search results for GPQA Diamond
AI & Enterprise
DeepSeek V4 seen having bigger impact than R1 on strong price performance
China\'s AI company DeepSeek has launched its new V4 models, drawing attention for offering open-source, near-frontier performance at a much lower price than Opus 4.7 or GPT-5.5. The V4 Pro and V4 Flash models were trained on about 33 trillion tokens and post benchmark results close to those rivals. Commentator Matthew Berman said pricing could pull U.S. companies toward DeepSeek, though geopolitical risks remain.
AI & Enterprise
Why Google Gemma 4 is in the spotlight as an open-source AI game changer
Google DeepMind has unveiled its open-source AI model Gemma 4, drawing a positive early response and renewed attention to the Gemma series. The lineup comes in four sizes, including edge models and larger variants with expanded context windows. Benchmarks cited show large gains in math, coding and science tests versus the prior version. Summit Pande highlights the efficiency of the 26B MoE model and points to Gemma 4’s adoption of the Apache 2.0 license.
AI & Enterprise
Upstage unveils in-house LLM Solar Pro 3 with 102 billion parameters
AI startup Upstage on Monday unveiled its in-house large language model, Solar Pro 3. The model has 102 billion parameters, more than triple the size of its predecessor, Solar Pro 2. The company said performance improved by more than twofold on key benchmarks including agent capability, coding and instruction following, while costs and throughput speed remained the same as the previous version. It also said reasoning improved using its SnapPO reinforcement learning technique.