Artificial Intelligence

Sep
26
Invide GLM-5.3 banner with a chart comparing GLM-5.3, Kimi K3, DeepSeek V4 Pro and Flash by score and cost per index task; below, GLM-5.3 leads Flash 42% to 33% on Terminal-Bench 4.0.

GLM-5.3 benchmarks, pricing and how it compares with Flash and rivals

GLM-5.3 keeps GLM-5.2's list price and matches Kimi K3's cost per Artificial Analysis index task. DeepSeek V4 Pro costs a third as much on the same index. Existing users must update requests that disable thinking.
8 min read
Sep
23
Banner comparing Opus 5.5 and GPT-6 Sol scores and task costs, with callouts for Luna’s $0.07 per Intelligence Index task and Grok 4.7’s more than doubled output tokens.

Opus 5.5, GPT-6 Sol, GPT-6 Luna and Grok 4.7 compared on cost

Opus 5.5's lower rates left max-effort task cost roughly level with Opus 5 in one independent test. Sol and Luna roughly halved their cost per task, while Grok 4.7 used more than twice the output tokens at unchanged rates.
4 min read
Sep
22
Four-step diagram for evaluating System One models: define answers, gather hard cases, test each approach, and score answers and probabilities.

What are System One models? Jev and its open-source alternatives

TypeSafe's Jev returns typed answers and probabilities instead of text. Laya, Nimble, Kev and OpenThai now offer open versions of the same idea. This guide covers how they work, their limits and test results, and when a plain classifier is the simpler option.
7 min read
Sep
17
GLM-5.3-Flash model guide comparing its $0.50-per-million output price with Opus 4.8 at $25, alongside questions about cost per finished task and its 306 GiB FP8 weights.

GLM-5.3-Flash costs up to 50x less than Opus. Your cost per finished task may not.

Given a billion tokens per vulnerability, GLM-5.3-Flash matched Claude Mythos Preview on ExploitBench at roughly 6% of the cost, while one production team found it fine for bounded work and not ready for their hardest tier.
10 min read
Sep
15
DeepSeek V4.1 Flash banner showing a 552B MoE backbone, 8B active on input, 16B on output, a 1M-token context, and $0.003/M off-peak cached input.

DeepSeek V4.1 Flash is a 552B model that bills like a smaller one

DeepSeek V4.1 Flash activates 8B parameters per input token and prices cached input from $0.003 per million. On Terminal-Bench 4.0, it scores 31.2 to Opus 5.0's 51.8.
6 min read
Sep
09
OpenAI Navier-Stokes banner contrasting two Codex data statements: access to specific user data is denied; de-identified data helping improve models is not ruled out.

OpenAI says its AI solved Navier-Stokes. Codex users have another question

Navier-Stokes, a Millennium Prize problem in fluid dynamics, is still listed as unsolved. OpenAI denies accessing any specific user data but cannot rule out de-identified data from the researchers' Codex use improving its models.
4 min read
Sep
04
GPT-6 Astra news banner showing an agent workflow flagged and stopped with HTTP 403 misalignment_policy_violation, with no general resume path.

GPT-6 Astra, OpenAI's first Critical cyber model, can stop your API job without asking

OpenAI says GPT-6 Astra can find unknown vulnerabilities and develop exploits with the right tools and access. Its API can terminate flagged agent work with a 403.
4 min read
Sep
02
Claude Fable 5.1 banner showing cache reads falling from $1 to $0.25 per million tokens and three breaking changes involving forced tool use, thinking blocks, and edited context.

Claude Fable 5.1 makes caching cheaper and breaks three API behaviors

Cache reads now cost $0.25 per million tokens, but forced tool use and preserved thinking require code changes before some agent workloads can migrate.
4 min read
Sep
01
ChatGPT banner showing 159.1 million monthly EU users above the 45 million VLOSE threshold.

The EU has designated ChatGPT a search engine because it searches the web

OpenAI reported 159.1 million monthly EU users, more than triple the DSA threshold. The designation decision explaining the classification is still unpublished.
3 min read
Aug
27
NVIDIA and Hugging Face logos alongside the reported, unconfirmed $12.9 billion deal and governance questions not addressed by their July open-weights letter.

NVIDIA's reported Hugging Face deal leaves platform governance unanswered

The reported $12.9 billion agreement remains unconfirmed. Hugging Face ran a sale process after reportedly rejecting a $500 million NVIDIA investment.
3 min read