Gemini 4 Hydrogen lists at $1.50 and $7.50 per million tokens, twice what Gemini 3.8 Flash charges today and exactly what Flash charges from 1 January. On Artificial Analysisβs index runs it still costs about a quarter as much per task, because it reads about half the input. The 2.5x speed claim holds up on outside data. The βavailable to everyoneβ line holds on the API and in Search, and less so in the free Gemini app.
A recap video of Lance Martinβs AI Engineer talk has been going around with seven takeaways and six slash commands for cutting what you pay Claude. I checked the commands and numbers against Anthropicβs changelog, pricing page and the skill files those commands run. The commands are real, under slightly different names. The 72% saving is a ceiling, measured on harnesses nobody had tuned. The cache rule of thumb in the talk is already out of date in the direction that makes it matter more, and the overnight hill climb is a little less sealed against overfitting than the recap says.
A new paper overloads vLLM for 107 seconds and then drops the load to 60% of capacity. With client retries on, the engine never recovers, and propagating cancellation, the obvious fix, still leaves 85% or more of later calls undelivered. The engineβs own prompt-token counter misses a third to half of the work. What brought service back was attaching retries to work already running, not killing it faster.
John Langfordβs farewell to Microsoftβs AI Frontiers group makes a pointed case for research over scaling: a ByteDance result suggests a hidden-state feedback trick can match a model twice its size, so βdo the mathβ on halving training and inference cost. I read both papers behind that sentence. The inference saving holds up. The training saving does not, by the papersβ own accounting.
An MIT group had GPT-6 Astra edit 0.016 percent of Qwen3.5-0.8Bβs weights, and MATH-500 went from 7.9 to 48.1 percent. Before the edit, 93 percent of the modelβs answers ran into the 32,768-token limit, in a setup its own model card warns about. The edit taught the model to stop, which is a real repair but a narrower one than better math, and it came from one of seven runs.
nesformer runs a Nintendoβs 6502 CPU inside an ordinary Qwen3.5 checkpoint. The headline says about 1GB of weights, but only 9.3 million of the 466.6 million parameters are nonzero, the picture is drawn outside the model, and it runs 150 to 290 times slower than the console. The part worth keeping is a different one: the weights are the machine, and the game lives in the modelβs state, which is somewhere nobody inspecting a model file would look.
Frontier API prices have fallen by up to 80% since June, Max Woolf writes. The headline cuts land on different parts of the bill. Fable 5.1βs cache reads are 75% cheaper while its input and output prices did not move, so a chat workload pays the same as before. On an agent workload I worked through, Opus 5.5 is about 28% cheaper than Opus 5, not 40 to 60.
Hazy Researchβs Next-Concept Modeling generates text in a fixed hierarchy of short-to-long code sequences instead of 128 denoising passes over the whole thing. The paper reports 10 to 31 times faster generation than diffusion baselines. The speedup is real in the tables, and the quality claims need reading one row at a time.
Between 18 September and 13 October, Googleβs own pages show a free-access line moving down the model ladder: the flagship is paid-only on the API, consumer-app users without a plan are routed mostly to the smallest model, and introductory paid prices carry dated step-ups. The free Flash tier on the API is still listed.
Googleβs EmbeddingGemma 2 is described as best in class for its size. As of 14 October 2026, its model card compares it with exactly one model, its predecessor, and reports a flat multilingual score. The comparisons that support the headline sit in three chart images in the launch post, and the license line says Apache 2.0 and also asks for adherence to a Gemma policy.
Qualcommβs CEO said AI companies want phones that can run at least a 100-billion-parameter model by 2028. It was a request, not a spec. Against a flagshipβs measured RAM and its derived memory bandwidth, the request hits a capacity wall first and a bandwidth wall second, and a sparse model only moves the second one.
The Financial Times says Nvidia is in early talks to buy Reflection AI, five days after Reflection announced its first open-weight model. As of 10 October the weights, the license file and the technical report are all still promises.
Samsungβs LittleBit paper compresses weights to 0.1 bits each. That number is counted one layer at a time. The memory saving for a whole model is about 21 times on Llama2-7B, not 160, and the quality cost at that budget is steep.
Claude Haiku 5.5 is priced 90% below Haiku 4.5 on every line for prompts up to 100,000 tokens and 50% below above it. Anthropic says it costs around 75% less on average. One independent benchmark puts the per-task drop at 25%. All three numbers are honest; they measure different things.
Aleph Alpha released Kolibri-1 today, a 78B-parameter model that activates 3.46B per token, billed as Germanyβs sovereign open model. Its own report says it leads the 3B-active class by about a point, and its headline efficiency chart prices decode throughput but not the 78 GB that has to sit in memory.
page 1 of 2 · 27 essays on Models, inference & pricing