Model Watch

Writing filed under Model Watch.

Fable 5, Sonnet 5, and GPT-5.6 raised the safety and routing bar

The 2026 frontier launches made raw capability harder to separate from safeguards, effort levels, pricing, and model routing.

July 10 20268 min read

#anthropic #openai #gpt-5.6 #claude-sonnet-5

Opir made guardrails look like classifiers again

Opir put safety filtering back in the small-model conversation with encoder guardrails for toxicity, jailbreaks, and policy categories.

May 29 20266 min read

#opir #classification #guardrails

Sonnet 4.5 and Gemini 3 made coding the frontier battleground

Anthropic and Google ended 2025 by treating coding, agents, and computer use as the proving ground for frontier models.

November 19 20256 min read

#anthropic #google #claude-sonnet-4.5 #gemini-3

Claude 4 and GPT-5 made agent work a maintenance problem

Claude 4 pushed long-running coding agents, then GPT-5 turned automatic reasoning into the default ChatGPT path.

August 8 20256 min read

#anthropic #openai #claude-4 #gpt-5

DeepSeek, Kimi, and MiniMax made open reasoning cheaper

The 2025 Chinese open-weight wave changed the cost story around reasoning, coding, and agentic model work.

July 15 20257 min read

#deepseek #kimi #minimax #reasoning models

Qwen 2.5 and Qwen3 made open models feel global

Alibaba's Qwen releases showed that the open-weight race was no longer centered on one country, one lab, or one model shape.

April 30 20256 min read

#qwen #alibaba #open weights

Spring 2025 split reasoning into products

Claude 3.7 Sonnet, Gemini 2.5 Pro, GPT-4.1, o3, and o4-mini made reasoning less like one capability and more like a set of product choices.

April 18 20257 min read

#openai #anthropic #google #reasoning-models

o1 and Gemini 2.0 pushed agents toward the default

OpenAI's o1 and Google's Gemini 2.0 moved the conversation from chat quality to reasoning time, tools, and agent-shaped product paths.

December 12 20247 min read

#openai #google #o1 #gemini-2.0

LFM made small models an architecture question

Liquid Foundation Models made small language models feel less like compressed LLMs and more like a different design bet.

October 1 20246 min read

#liquid ai #lfm #small language models

Llama 3 and 3.1 made open weight the default baseline

Meta's 2024 Llama releases changed the question from whether to test open weights to where they fit in the stack.

July 24 20245 min read

#meta #llama #open weights

GPT-4o and Claude 3.5 Sonnet compressed the frontier

The May and June 2024 launches made frontier quality feel faster, cheaper, more visual, and harder to rank with one benchmark.

June 22 20246 min read

#openai #anthropic #gpt-4o #claude-3.5-sonnet

GLiNER turned entity extraction into a small-model job

GLiNER was a useful reminder that not every language task needs a decoder LLM pretending to be a classifier.

May 10 20247 min read

#gliner #ner #classification

Gemma and Phi made small models feel serious again

Google Gemma and Microsoft Phi-3 pushed small open models into the product conversation instead of leaving them as demo hardware toys.

April 23 20247 min read

#gemma #phi #small language models

Claude 3 gave the model picker real tradeoffs

Claude 3 turned Anthropic's lineup into a practical choice between speed, price, and capability instead of one vague assistant.

March 5 20247 min read

#anthropic #claude-3 #model-watch

Gemini 1.5 made long context the main event

Gemini 1.5 Pro changed the Google story from catch-up to a specific advantage: very large context across mixed media.

February 16 20247 min read

#google #gemini-1.5 #long-context

Mixtral made open weights feel usefully weird

Mixtral 8x7B made sparse open-weight models feel like a practical path instead of a research curiosity.

December 12 20235 min read

#mistral #mixtral #open weights

Gemini 1 arrived with a multimodal reset

Google's Gemini 1.0 launch made the next model fight feel multimodal, tiered, and product-shaped.

December 7 20236 min read

#google #gemini #model-watch

GPT-4 Turbo made context a product budget

OpenAI's DevDay launch made bigger context windows feel like something product teams had to price, test, and explain.

November 7 20236 min read

#openai #gpt-4-turbo #model-watch