| OpenAI |
o1-preview |
2024-09-12 |
128K |
Preview of the o-series reasoning models. It runs an internal chain of thought before answering and was clearly stronger than then-current GPT-4o on math, coding, and science. |
| OpenAI |
o1-mini |
2024-09-12 |
128K |
A lighter reasoning version of o1, focused on STEM and coding. Lower cost and faster, for cases that need some reasoning but not the strongest results. |
| DeepSeek |
DeepSeek-V2.5 |
2024-09-05 |
128K |
Merged V2-0628 and Coder-V2-0724 so general chat and coding share one model. |
| xAI |
Grok-2 |
2024-08-13 |
128K |
A full-generation Grok upgrade with stronger coding and writing, plus realtime X search and image generation. |
| OpenAI |
GPT-4o mini |
2024-07-18 |
128K |
A cost-efficient small multimodal version of GPT-4o. It replaced GPT-3.5 Turbo in ChatGPT and fits high-volume lightweight tasks. |
| Anthropic |
Claude 3.5 Sonnet |
2024-06-20 |
200K |
A breakout Sonnet generation. Coding, vision, and tool use far surpassed Claude 3 Opus, with excellent value. |
| DeepSeek |
DeepSeek-Coder-V2 |
2024-06-17 |
128K |
A code-specialized model based on V2. Covers more programming languages and is strong at long-context repo Q&A. |
| Qwen |
Qwen2 |
2024-06-06 |
131072 |
The third Qwen generation, stronger in multilingual and long-context work. 7B and above generally support 128K. |
| Zhipu GLM |
GLM-4-9B |
2024-06-05 |
128K |
A compact open-source GLM-4, plus an experimental 1M version, for local and private deployment. |
| Gemini |
Gemini 1.5 Flash |
2024-05-14 |
1M |
A high-speed model distilled from 1.5 Pro. Low price and low latency, for high concurrency and realtime apps. |
| OpenAI |
GPT-4o |
2024-05-13 |
128K |
The first native omni multimodal model, handling text, image, and audio in one system with lower latency. It was ChatGPT's default model for a time. |
| DeepSeek |
DeepSeek-V2 |
2024-05-06 |
128K |
An MLA-architecture MoE. Performance approached then-current mid-tier closed models at a very low price, and it directly drove a domestic API price war. |
| xAI |
Grok-1.5 |
2024-03-28 |
128K |
Clearly stronger than Grok-1 on reasoning and instruction following. Context grew to 128K, and vision (Grok-1.5V) followed soon after. |
| Kimi |
moonshot-v1-128k |
2024-03-20 |
128K |
An early flagship on the Kimi open platform. 128K context, suited to Chinese long-document Q&A and summarization. |
| Anthropic |
Claude 3 Haiku |
2024-03-13 |
200K |
The fastest and cheapest Claude 3 tier, nearly instant. Fits classification, extraction, customer support, and other high-concurrency work. |
| Anthropic |
Claude 3 Opus |
2024-03-04 |
200K |
The Claude 3 flagship. At the time it matched GPT-4-class reasoning, math, and vision, with 200K context for complex analysis. |
| Anthropic |
Claude 3 Sonnet |
2024-03-04 |
200K |
The Claude 3 balanced tier, trading off intelligence and speed. It was the most commonly deployed Sonnet for enterprise use. |
| Gemini |
Gemini 1.5 Pro |
2024-02-15 |
1M |
The first to bring production context to 1 million (later experimented to 2 million). Long video, codebases, and extra-long documents are strengths. |
| Gemini |
Gemini 1.0 Ultra |
2024-02-08 |
32768 |
The strongest 1.0 tier, launched with Gemini Advanced. It targeted then-current GPT-4-class reasoning and multimodality. |
| Qwen |
Qwen1.5 |
2024-02-04 |
32768 |
A full generation aligned to ChatML. Sizes from 0.5B to 110B MoE, and the ecosystem started to take shape. |
