Release cadence by provider
The most recent model release date for each provider, alongside lineup size and the strongest model by average score.
What the data shows
Release dates cluster around a few periods, but the index includes models from 2023 through 2026. A recent release is not automatically stronger, and an older model can still be the best value for a narrow task.
Use the latest release date as a freshness signal, not a quality score.
How to read it
- Check the release date on the model page.
- Compare the latest model with the previous generation.
- Watch for price changes after a new release.
- Do not migrate without an evaluation.
Release cadence by provider
| Provider | Models | Latest release | Best model |
|---|---|---|---|
| 01.AI | 14 | 2026-10-27 | Yi-Lightning |
| Anthropic | 32 | 2026-10-26 | Claude Opus 4.5 Pro |
| OpenAI | 40 | 2026-10-22 | GPT-5.2 Pro |
| Mistral AI | 35 | 2026-10-22 | Mistral Large 3 |
| Cohere | 21 | 2026-10-21 | Command A Pro |
| xAI | 14 | 2026-10-16 | Grok 4 Reasoning |
| Moonshot AI | 14 | 2026-10-11 | Kimi K2 |
| Google DeepMind | 31 | 2026-10-09 | Gemini 3 Pro Max |
| Zhipu AI | 14 | 2026-10-07 | GLM-4.6 Instruct |
| DeepSeek | 21 | 2026-10-01 | DeepSeek-V3.2 |
| MiniMax | 14 | 2026-09-19 | MiniMax-M2 |
| Baidu | 14 | 2026-09-08 | ERNIE 4.5 Max |
| Alibaba Cloud | 28 | 2026-09-06 | Qwen3-Max Pro |
| Meta AI | 28 | 2026-08-28 | Llama 4 Maverick Instruct |
| Tencent | 14 | 2026-08-16 | Hunyuan-Turbo Max |
| ByteDance | 14 | 2026-08-12 | Doubao 1.5 Pro Reasoning |
| Together AI | 14 | 2026-08-01 | Llama 3.1 Turbo |
| Amazon | 14 | 2026-07-07 | Nova Pro Max |
| Perplexity | 14 | 2026-07-04 | Sonar Pro Max |
| Reka AI | 14 | 2026-02-10 | Reka Core Pro |
| Cerebras | 7 | 2025-12-27 | Llama 3.1 Cerebras Instruct |
| Databricks | 7 | 2025-12-23 | DBRX |
| Snowflake | 7 | 2025-12-07 | Arctic Instruct |
| Stability AI | 7 | 2025-12-04 | StableLM 2 Instruct |
| Groq | 14 | 2025-11-27 | Llama 3.3 Groq |
| EleutherAI | 14 | 2025-11-26 | GPT-NeoX Instruct |
| Lambda | 7 | 2025-10-23 | Llama 3 Lambda Instruct |
| Fireworks AI | 14 | 2025-10-20 | Firefunction Max |
| Hugging Face | 14 | 2025-10-18 | SmolLM3 Instruct |
| Allen Institute for AI | 14 | 2025-10-01 | OLMo 2 Instruct |
| OpenRouter | 14 | 2025-09-18 | Fusion Max |
| Modal | 7 | 2025-09-17 | Llama 3 Modal Instruct |
| NVIDIA | 14 | 2025-09-12 | Nemotron 4 |
| DeepInfra | 7 | 2025-08-23 | Llama 3 DeepInfra Instruct |
| Baseten | 7 | 2025-08-16 | Llama 3 Baseten Instruct |
| Microsoft | 14 | 2025-08-11 | Phi-4 |
| Replicate | 7 | 2025-07-23 | Llama 3 Replicate |
| Anyscale | 7 | 2025-06-10 | Llama 3 Anyscale Instruct |
| Novita AI | 7 | 2025-04-12 | Llama 3 Novita Instruct |
| Hyperbolic | 7 | 2025-02-11 | Llama 3 Hyperbolic Instruct |
Methodology
This story uses the AI & LLM Index dataset: 600 models, 40 providers and 12 benchmarks. Prices are listed per million tokens. Scores are modelled estimates on a 0–100 scale and are not official benchmark results. Figures are rounded for readability.
Related
FAQ
What does the Release cadence by provider data show?
The most recent model release date for each provider, alongside lineup size and the strongest model by average score.
Are the numbers official?
No. Prices and scores are compiled for this demo index. Benchmark values are modelled estimates, not official results. Verify figures with the provider or benchmark owner before making decisions.
How should I use this ranking?
Use it to form a shortlist. Check the specific model pages, compare price and context, and test the shortlist on your own examples before committing.