query://ai-newsJul 8, 2026 · 3 min read
Grok 4.5 loses on benchmarks and might win anyway
Grok 4.5 trails Fable 5 and GPT 5.5 on the hardest coding tests. It is also a quarter of Opus's price and burns 4x fewer tokens, which might matter more.
query://tags/llms
4 stories tagged llms
query://ai-newsJul 8, 2026 · 3 min read
Grok 4.5 trails Fable 5 and GPT 5.5 on the hardest coding tests. It is also a quarter of Opus's price and burns 4x fewer tokens, which might matter more.
Jul 6, 2026 · 3 min read
A proxy renders bulky prompt text as images to dodge token costs, cutting AI bills by up to 70 percent. Here is how the trick works, and where it quietly breaks.
Jun 7, 2026 · 3 min read
DeepSeek V4 Pro now starts at $0.003625 per million tokens, runs a 1M context on Huawei silicon, and just turned a promotion into the new normal. What the price war means for everyone's API bill.
Jun 4, 2026 · 3 min read
At Build 2026, Microsoft unveiled seven in-house MAI models trained without OpenAI, put MAI-Code-1-Flash in the Copilot picker, and signed the largest government contract in its history.