{"slug":"grok-4-5-launch","label":"Grok 4.5 launch","item_count":2,"day_count":2,"source_count":2,"first_seen":"2026-07-09T06:05:41+00:00","last_updated":"2026-07-10T02:56:40+00:00","via_scout":true,"generated_at":"2026-07-29T15:14:44.262987+00:00","sources":["latent_space","search_agent_engineering_news"],"days":[{"date":"2026-07-09","items":[{"title":"[AINews] SpaceXAI launches Grok 4.5, first Opus-class model post Cursor acquisition","url":"https://www.latent.space/p/ainews-spacexai-launches-grok-45","source":"latent_space","type":"news","summary_1line":"SpaceXAI continues to move faster than any other frontier lab on earth.","sid":"01ed778e69e3f41e","published":"2026-07-09T06:05:41+00:00","editor_note":"AINews/latent_space reports SpaceXAI launching Grok 4.5, its first Opus-class model since acquiring Cursor."}]},{"date":"2026-07-10","items":[{"title":"Grok 4.5 Cuts Coding-Agent Cost 80%: Near-Frontier Speed, Higher Hallucinations - Tech Times","url":"https://news.google.com/rss/articles/CBMiygFBVV95cUxQb095YTlFb1FrQjJBc25tY2J2Q3htSnR0WUxmekE0NUdhRXZxWFNCWnR4NWE1emZFU0tjdDNoeWJfZENfTUM1bWtfbEhJZFdKUVhESzV6cXplY0o5R2l6V3IxWVpweGhFbkxQRXF6Qi00VFhhb0lSX1BvN3VySGtyNGhnTWJDdEpCTkFLcDY2cWJ5Qy1XTzlfWVYySGpocGlIN0tITXBQaTR4RUFCVHJCQi1rLUc4UEZoWkNxT2xKNURqQm40LXhmN09n?oc=5","source":"search_agent_engineering_news","type":"news","summary_1line":"Grok 4.5 Cuts Coding-Agent Cost 80%: Near-Frontier Speed, Higher Hallucinations Tech Times","why_it_matters":"Matches feed focus: agent.","sid":"68e97756211ddc61","published":"2026-07-10T02:56:40+00:00","editor_note":"Tech Times publishes the first hands-on benchmark: 80% lower coding-agent cost at near-frontier speed, but more hallucinations than rivals."}]}],"editorial":{"tldr":"SpaceXAI shipped Grok 4.5 on Jul 9, its first Opus-class model since acquiring Cursor. Coverage framed it as the fastest-moving frontier lab shipping yet another release.","stale":false,"whats_new":"A day-two benchmark write-up found Grok 4.5 cuts coding-agent cost 80% at near-frontier speed, but with a higher hallucination rate than rival models.","why_it_matters":"Grok 4.5's price/speed profile makes it worth a coding-agent cost trial, but its higher hallucination rate means it needs stricter output verification than Opus-class alternatives before production use.","take_for_builders":"If you're evaluating coding-agent cost, benchmark Grok 4.5 against your current model on your own tasks before switching — the 80% cost cut is real per Tech Times, but weigh it against the reported higher hallucination rate for your specific verification budget.","status":{"state":"Shipped · hands-on benchmark published","tone":"launch","changed":"2026-07-10","detail":"Launched Jul 9; a Jul 10 benchmark write-up is the only independent read on cost, speed, and reliability so far."},"beats":[{"kicker":"LAUNCH","tone":"launch","headline":"SpaceXAI ships Grok 4.5, its first Opus-class model post Cursor acquisition","summary":"AINews/latent_space frames the release as SpaceXAI continuing to out-ship every other frontier lab.","sids":["01ed778e69e3f41e"]},{"kicker":"NOW","tone":"now","headline":"Hands-on benchmark: 80% cheaper coding-agent runs, near-frontier speed, more hallucinations","summary":"Tech Times' day-two benchmark is the first independent read on Grok 4.5's coding-agent economics and reliability tradeoff.","sids":["68e97756211ddc61"]}],"open_questions":["Does the 80% coding-agent cost cut hold up across benchmarks beyond Tech Times' initial write-up?","How much does the higher hallucination rate cost in practice for agentic coding workflows — does it require extra verification steps that erode the price advantage?","Does SpaceXAI publish its own benchmark numbers, or does independent testing stay the only source?"],"generated_at":"2026-07-29T15:25:00Z"}}