AI Daily · Nov 13, 2026: Google releases Gemini 4 Hydrogen: 2.5x faster, priced at $1.50/$7.50
Gemini 4 Hydrogen is rolling out globally with 2.5x faster task completion than 3.8 Flash, a 1M-token context, four thinking levels, and pricing at $1.50/$7.50 per million input/output tokens; free users reach it via auto model selection.
Pick a date
-
Google releases Gemini 4 Hydrogen with 2.5x faster task completion
Gemini 4 Hydrogen targets everyday tasks, coding and agentic workflows, finishing tasks 2.5x faster than 3.8 Flash with four thinking levels so users pay only for the reasoning depth needed. It is live in the Gemini API, Gemini for business, the Gemini app, Google Search and Antigravity, with free users reaching it via auto model selection.
-
Gemini 4 Hydrogen priced at $1.50 input and $7.50 output
Gemini 4 Hydrogen costs $1.50 per million input tokens and $7.50 per million output tokens, with a 1M-token context window; LiteLLM notes it surpasses Argon on the AA Intelligence Index and is 2.5x faster than 3.8 Flash.
-
DeepSeek opens App capabilities via API including OCR, ASR and TTS
DeepSeek now exposes the App's search, file, link, safety and voice features through its API, with standalone OCR, ASR and TTS endpoints. Extended capabilities carry tool usage fees on top of standard token charges.
-
Perplexity open-sources Decider v1.3 decision models, free on OpenRouter for a week
pplx-decider-v1.3 and pplx-decider-flash-v1.3 are open-weight and free on OpenRouter for one week; v1.3 adds a "locate" feature that returns matching text spans with character offsets.
-
Thinking Machines' Inkling 2 Small 594B MoE model now available on Ollama
Ollama added Inkling 2 Small from Thinking Machines, a 594B mixture-of-experts model with a 1M-token context window and zero data retention, targeting OCR, grounding, infographic interpretation and quantitative reasoning.
-
Inkling 2 Small weights released on Hugging Face under Apache 2.0
Thinking Machines published the weights of Inkling 2 Small on Hugging Face under the Apache 2.0 license, allowing download and local customization, following safety testing with partners.
-
B.AI cuts GLM-5.3-Flash discount to 40% off in model promo update
B.AI's discount matrix now offers GLM-5.3-Flash at 40% off (down from 30%), Hy4 at 50% off until Dec 31 2026, DeepSeek-V4.1-Flash at 40% off and Qwen3.8-Flash at 20% off. Hy4 adds image input with a 1M-token context; GLM-5.3-Flash is a 320B native multimodal model.
-
Command Code GOAT plan offers $100 of DeepSeek V4.1 Flash usage for $10/month
Command Code's GOAT plan costs $10 per month and includes $100 of usage for running the full-weight DeepSeek V4.1 Flash model, pitched as faster and cheaper.
-
llama.cpp releases v1.0.0 as first fully stable major version
llama.cpp shipped v1.0.0, described as the first major fully stable release of the widely used library for running AI models locally, including on phones; the team pledged continued coverage of every platform and model.
-
Claude Meetings beta rolls out to Team and Enterprise plans
Anthropic is rolling out Claude Meetings in beta on Team and Enterprise plans; Claude joins video calls and builds docs, whiteboards and other artifacts alongside users in real time. It starts with Google Meet support, with notes, action items and transcripts appearing on the meeting page in Claude.