OpenAI reviewed the SWE-Bench Pro AI coding test and found roughly 30% of its tasks are flawed, leading the company to withdraw its endorsement. The problems include overly strict or vague task requir…
SpaceXAI launched Grok 4.5, a coding-focused AI model priced at $2 per million input tokens and $6 per million output tokens, aiming to reduce costs for AI-assisted software development. The model is …
XAI released Grok 4.5 on 8 July 2026, priced at $2 per million input tokens and $6 per million output tokens, making it roughly five times cheaper than competitors like Claude Fable 5 and GPT-5.5. The…
Artificial Analysis launched AutomationBench-AA on July 7, 2026, an independent leaderboard for Zapier's AutomationBench that grades AI agents on final-state outcomes of simulated SaaS workflows. The …
SpaceXAI's Grok 4.5 ranked fourth on Artificial Analysis' Intelligence Index, scoring 54 behind Fable 5, GPT-5.5, and Opus 4.8. The model showed strong cost efficiency at $0.31 per task and competitiv…
SambaNova combined four Nvidia H200 GPUs with 16 SN50 RDU chips to achieve 763 tokens per second on MiniMax M2.7 at a 10,000-token context, separating prefill and decode stages to accelerate inference…
Anthropic's Claude Fable 5 tops six new industry benchmarks from Artificial Analysis in finance, law, and medicine, but costs over 100 times more per task than DeepSeek V4 Pro despite a narrow perform…
Artificial Analysis launched Harvey LAB-AA, an independent benchmark testing AI models on 120 legal tasks across 24 practice areas. Claude Fable 5 led with a 14.2% all-pass rate, though results highli…
Anthropic cut Sonnet 5's price to $2 per million tokens, but Artificial Analysis measured the actual per-task cost at $2.29, exceeding Opus 4.8's cost. The discrepancy stems from the effort dial setti…
Z.ai's open-weight GLM-5.2 model, released in June 2026, approaches proprietary system performance on coding and agentic tasks while compressing inference costs, according to benchmarks and developer …
Model subsidies are ending as flat-rate AI subscriptions prove unsustainable for agentic workloads, which consume up to 1,000 times more tokens than standard queries. Artificial Analysis' AA-Briefcase…
Wagtail developer Thibaud Colas published a guide comparing open weight AI models and inference providers, emphasizing transparency and decoupled infrastructure as key advantages. The post outlines cr…
Anthropic's Claude Sonnet 5 achieves a score of 53 on the Artificial Analysis Intelligence Index, matching GPT-5.5 with high reasoning but costing $2.29 per task—15% more than Opus 4.8—due to increase…
Google DeepMind launched the Nano Banana 2 Lite, officially branded as Gemini 3.1 Flash Lite Image, on February 26, 2026, ranking fifth in text-to-image leaderboards with an Elo score of 1,255. The mo…
Cerebras announced that Gemma 4 31B, Google DeepMind's open-weight multimodal model, is now available on its platform, achieving over 1,800 tokens per second—35x faster than typical GPU endpoints. The…
European AI experts push back against the narrative of permanent American dominance, arguing that the gap between US proprietary models and open-source alternatives is narrowing. The persistence of th…
Z.ai released GLM-5.2, an open-weight text-only model with a 1 million token context window, priced at one-fifth the output cost of Claude Opus. Independent tests by James Daniel Whitford at TechStack…
A viral chart predicting open-weights AI models will catch closed frontier models by December 2026 is misleading, as analysis across 18 benchmarks shows the gap varies by task and is not uniformly shr…
Analysis of 18 benchmarks suggests open-source LLMs may catch up to closed-source models by December 3, 2026, based on one index, but the average gap across all benchmarks remains flat at about 5 mont…
Artificial Analysis launched the Speech to Speech Index, a new metric combining Big Bench Audio, Full Duplex Bench, and τ-Voice to evaluate native Speech to Speech models. OpenAI's GPT-Realtime-2 (Hig…