Platform

Claude Opus 4.8 cuts fast-mode pricing by two-thirds, adds reasoning sliders

Anthropic released Claude Opus 4.8 on May 28, keeping base pricing flat while slashing fast-inference costs and adding granular reasoning-length controls.

ByAlex Sokoloff·May 29, 2026

Claude Opus 4.8 cuts fast-mode pricing by two-thirds, adds reasoning sliders

Anthropic released Claude Opus 4.8 on May 28 with three headline changes: granular reasoning controls that let users dial inference depth up or down, a fast mode now priced at one-third the cost of the previous fast tier, and what the company calls "improved honesty" — a claim that the model will hallucinate less often. The base pricing for Opus 4.8 remains unchanged from the prior version.

The fast mode shift is the most tangible cost win. Anthropic's fast inference option already delivered 2.5× faster generation; the new pricing drops the premium from 6× the base rate to 2× — a three-fold reduction that makes the fast tier viable for higher-volume workloads. The reasoning controls mirror the slider OpenAI added to ChatGPT, giving practitioners fine-grained control over how much compute the model spends on chain-of-thought steps before answering.

What stands out

More in Platform

01

Granular reasoning controls — Users can now adjust inference depth on a per-request basis, matching the UX pattern ChatGPT introduced. The feature is live in the API and web UI.

02Fast mode repricing — The fast inference tier now costs 2× the base rate instead of 6×, while still delivering 2.5× faster generation. That puts the speed premium at roughly 0.8× per token of wall-clock time saved — a meaningful improvement for latency-sensitive applications.

03Honesty claims — Anthropic's release notes call out "improved honesty" as a core upgrade in Opus 4.8, positioning the model as less prone to confabulation. No benchmark numbers or eval methodology were shared in the announcement.

04Next tier teased — The company said it plans to ship "a new class of models with even higher intelligence than Opus" in the coming weeks, suggesting a tier above the current flagship is already in late testing.

05Pricing stability — Base rates for Opus 4.8 match the prior version, holding the line on the flagship tier while cutting costs in the fast lane.

Research

NSFW

Platform

Confident agents hijack multi-LLM debates, arXiv study shows

Researchers model multi-agent LLM systems as mixtures of experts governed by social-influence dynamics, where confident agents dominate group decisions regardless of correctness.

May 31, 2026

IndustryNSFW

Google's AI search leaves brands blind to how they're described

Google I/O confirmed AI-generated answers now dominate search results, upending SEO strategies built around traditional link rankings and leaving most brands without insight into how AI systems describe them.

May 30, 2026

ReleasesNSFWPlatform

Mero-Artemis-31B-v0.3.1-heretic: uncensored 31B multimodal merge on Gemma4

sh0ck0r released Mero-Artemis-31B-v0.3.1-heretic, a 31-billion-parameter uncensored image-text-to-text model built on Gemma4 and available in safetensors format on HuggingFace.

May 30, 2026

PlatformNSFW

YouTube moves AI labels to top of Shorts, auto-detects synthetic content

YouTube is relocating AI disclosure badges to more visible positions on Shorts and long-form videos and will begin automatically identifying AI-generated content, the company announced this week.

May 30, 2026

ReleasesNSFW

Llama 3.2 3B uncensored GGUF quantization debuts for CPU inference

Jadzblaze99 published quantized GGUF weights for an uncensored fine-tune of Meta's Llama 3.2 3B Instruct model, enabling CPU and low-VRAM deployment of a 3-billion-parameter conversational model without safety filters.

May 30, 2026