Topic
Per-token price levels and cuts for hosted model inference from major labs.
Subquadratic launched SubQ Snap at $0.05 per million input tokens with a 384K-token window, the sixth hosted decision model in about three weeks. No outside lab has run it, so every score in the launch is a claim about Subquadratic's own test set.
Reality
Elon Musk said SpaceX's Grok Bot agent will send each task to the best back-end model, Anthropic's Claude Opus 5.5 included. The policy moves the engineering work into routing, and into the tests that decide what 'best' means for each task.
Reality
Cloudflare's Clef-omni takes audio and video alongside text and images and returns schema-bound decisions in one API call. Runtimewire reported that the same release cut hosted Clef-flash's context window from 64,000 to 24,000 tokens, so longer prompts now go to Clef or to self-hosted weights.
Perspective Coverage
3 publishersReality
Antseed launched a peer-to-peer inference marketplace that it says cuts frontier-model costs by up to 97% below official API prices. The figure is Antseed's own, so the first job for a team is to check it on work whose prompts it can afford to send to a stranger.
Reality
Ramp's AI Index shows US corporate token use up about 50% since July while AI spending fell, as effective token prices dropped 41%. Usage has not grown enough to cover the price cut, so competition between OpenAI and Anthropic has so far moved revenue from the two sellers to their customers.
Reality
Diogo Almeida's TypeSafe put Jev into early access on 15 September with $40m from DCVC, and it sells a calibrated confidence number on every answer as the thing that makes automation possible, with the evaluations behind that claim built in-house.
Reality
Epoch AI puts the price of answering a GPQA Diamond question at a fixed accuracy bar falling about 13 times a year, faster than compute under Moore's Law. Buyers get that discount only by moving to newer models, so the part of a stack that has to switch cheaply is the evaluation that qualifies each one.
Reality
Featherless open-sourced Simple Jev, a library that has open models pick fixed labels or yes/no answers, with hosting from $0.03 per million input tokens. CEO Eugene Cheah pitches it for classification work teams now pay frontier-model prices to run.
Reality
Futurum's report, sponsored by neocloud QumulusAI, says agentic tasks can burn 10 to 100 times the tokens of a single inference call. Because the meter bills every step, the agents that win real users are the ones that run through their budgets first.
Reality
TypeSafe AI is in talks to raise over $1 billion at a valuation above $10 billion, 50 times its seed price from two weeks earlier. Buyers at that price are paying for launch-week usage and an investor's profit claim, with no revenue figure published.
Reality
Anthropic and OpenAI each launched a model on Tuesday that repackages capability their flagships already had. The faster release calendar behind those launches is mostly a pricing story for the people who buy them.
Reality
Artificial Analysis scores the new Xiaomi model first among open weights and twelve points behind Claude Opus 5.5, and the cheaper Flash tier is the one an operator should put in front of a real queue.
Reality
Version 2.0 arrives at the same price as version 1.0. The accuracy gain SpaceXAI reports is concentrated on short multilingual commands. The evaluation sets behind the claim are drawn from production support calls.
Reality
Google's September 15 release pairs a fast speech-to-speech model with one that narrates its own reasoning aloud while tools run. Clients now have to read interactionStatus to know when a turn is actually over.
Reality
Ramp's corporate spending data puts the effective US business price at 68 cents a million tokens, down from $1.15 in March, and the customers who supply most enterprise AI revenue trimmed per-employee spend as it fell.
Reality
Steady list prices have pushed the variance into workload design. BCG's benchmark finds the cheapest cloud changes with the workload, while the decision it calls costliest over time falls outside what the benchmark measures.
Reality
Chinese banks and telcos are packaging inference the way airlines package miles, though the analyst closest to the trend calls the consumer bundles a supply-led experiment whose users never see a balance.
Reality
VLM Run's gateway lets developers swap open-weight document models without touching their pipelines. The routing argument is sound; the pricing claim arrives without accuracy data.
Reality
Per-token prices fell about 200x since GPT-4's launch while US enterprise AI spend tripled to $37 billion. Tokenizer variance, reasoning tokens and tier discounts are where the bill diverges.
Reality
Silicon Data figures reported by the FT put the one-month drop at close to 25%. The pressure is coming from DeepSeek and Moonshot, not from OpenAI and Anthropic fighting each other.
Reality