AI UpdatesToday
Track AI model updates and LLM releases in real time. Version launches, API changes, and notable improvements across GPT, Claude, Gemini, Llama, and 500+ language models.
Loading updates...
Last 10
AI Model Releases
Track all LLM releases and version updates
Jun 12, 2026· 1 release
Jun 10, 2026· 1 release
Jun 9, 2026· 1 release
Jun 5, 2026· 1 release
Jun 2, 2026· 2 releases
Jun 1, 2026· 1 release
May 28, 2026· 1 release
1–8 of 316
Last 30 days
Models that got better
Sigma-normalized vs. each model's baseline
Gemini 3.1 Flash Image
Google·3mo old
+2.09σ
now +4.95σ
106 votes·1 arena
Claude Opus 4.7
Anthropic·1mo old
+1.28σ
now +2.01σ
337 votes·2 arenas
Claude Opus 4.6
Anthropic·4mo old
+0.65σ
now +4.07σ
127 votes·1 arena
How is this measured?⌄
For each model we reconstruct daily TrueSkill conservative ratings per arena from match-level vote outcomes, then compute a baseline from the first 21 days of activity (after a 3-day warm-up). The Quality Index is the sigma-normalized deviation from that baseline, weighted across arenas. Change shown is the difference between today and 30 days ago. A swing of ±0.5σ is noticeable; ±1σ is significant.
Weekly brief
The model releases, benchmark shifts, and analysis worth your week — in one email.
Open weights
Open Source AI Updates
Recent open-weight model releases with permissive licenses
Landscape
Open Source LLM Landscape
Open source LLM news has become increasingly important as open-weight models transform the AI landscape. Stay updated with open source LLM updates today covering models like Llama 3, Mistral, Qwen, and DeepSeek—now rivaling proprietary alternatives on many benchmarks while providing flexibility to fine-tune, self-host, and customize for specific domains.
Our open-source LLM coverage includes licensing terms (Apache 2.0, MIT, or custom licenses), parameter count affecting LLM inference costs, quantization support for efficient deployment, and the community ecosystem of fine-tuned variants and LLM tools.
Primer
Understanding LLM Versioning
AI model versioning follows patterns that help developers understand capabilities and stability. Major versions (GPT-3 → GPT-4, Claude 2 → Claude 3) indicate significant capability improvements and may require prompt adjustments. Minor updates (GPT-4 → GPT-4 Turbo) offer performance optimizations, cost reductions, or context window expansions while maintaining compatibility.
Organizations use various naming conventions: OpenAI uses dated snapshots (gpt-4-0613), Anthropic uses descriptive tiers (Claude 3.5 Sonnet), and Google uses generation markers (Gemini 1.5 Pro). Understanding these patterns helps you make informed decisions about when to upgrade and how to manage deprecations.
Labs
Active AI Organizations
Track model releases from leading AI labs
Moonshot AI
9 models•Updated 2d ago
Recent Models
1Kimi K2.7 Code
2Kimi K2.6
3Kimi K2.5
3 models in 6 mo
JFMAMJ
47 models•Updated 4d ago
Recent Models
1DiffusionGemma 26B-A4B
2Gemma 4 12B
3Gemini 3.5 Flash
9 models in 6 mo
JFMAMJ
Anthropic
19 models•Updated 5d ago
Recent Models
1Claude Fable 5
2Claude Opus 4.8
3Claude Opus 4.7
6 models in 6 mo
JFMAMJ
Unisound
1 models•Updated 1w ago
Recent Models
1U2
1 models in 6 mo
JFMAMJ
Microsoft
12 models•Updated 1w ago
Recent Models
1MAI-Thinking-1
2MAI-Code-1-Flash
3MAI-Image-2.5
2 models in 6 mo
JFMAMJ
MiniMax
14 models•Updated 1w ago
Recent Models
1MiniMax M3
2MiniMax M2.7
3MiniMax M2.5
3 models in 6 mo
JFMAMJ
Alibaba Cloud / Qwen Team
53 models•Updated 3w ago
Recent Models
1Qwen3.7 Max
2Qwen3.6-27B
3Qwen3.6-35B-A3B
12 models in 6 mo
JFMAMJ
xAI
24 models•Updated 1mo ago
Recent Models
1Grok 4.3
2Grok-4.20 Beta Non-Reasoning
3Grok-4.20 Multi-Agent Beta
4 models in 6 mo
JFMAMJ
OpenAI
59 models•Updated 1mo ago
Recent Models
1GPT-5.5 Instant
2GPT-5.5
3GPT-5.5 Pro
9 models in 6 mo
JFMAMJ
Side by side
Compare AI models
Free head-to-head playgrounds across image, video, website, game and chat modalities.
Trends
The Pace of AI Development
The AI industry is releasing new models at an unprecedented rate. We track 316+ model releases across major organizations. Capabilities that seemed cutting-edge months ago are now baseline expectations.
Key trends include reasoning models (OpenAI o1, DeepSeek-R1) trading speed for accuracy, multimodal capabilities becoming standard across frontier models, and efficiency improvements delivering GPT-4-level performance at dramatically lower costs.
50+ organizations·15+ providers·Updated daily
Inference
API Provider Updates
Pricing, latency, and feature updates from inference providers
Replicate
99 active models·replicate.com/
Input price
—
Max tokens
10K
Fal
70 active models·fal.ai
Input price
—
Max tokens
8K
OpenAI
63 active models·openai.com
Input price
$0.10–$10.00
Max tokens
1.1M
Popular Models
GPT-5.5$500.00/M
GPT-5.5$500.00/M
46 active models·ai.google.dev
Input price
$0.25–$2.50
Max tokens
1.0M
Popular Models
Gemini 3.5 Flash$150.00/M
Gemini 3.5 Flash$150.00/M
Novita
42 active models·novita.ai/
Input price
$0.08–$2.00
Max tokens
1.0M
Popular Models
Kimi K2.7 Code$95.00/M
Kimi K2.7 Code$95.00/M
xAI
22 active models·docs.x.ai
Input price
$0.20–$3.00
Max tokens
2.0M
Popular Models
Grok-4 Fast Non-Reasoning$20.00/M
Grok-4 Fast Non-Reasoning$20.00/M
DeepInfra
20 active models·deepinfra.com/
Input price
$0.06–$1.74
Max tokens
1.0M
Popular Models
MiMo-V2.5-Pro$100.00/M
MiMo-V2.5$40.00/M
Anthropic
18 active models·anthropic.com
Input price
$1.00–$10.00
Max tokens
1.0M
Popular Models
Claude Fable 5$1000.00/M
Claude Fable 5$1000.00/M
MiniMax
12 active models·platform.minimax.io
Input price
$0.30–$0.60
Max tokens
1.0M
Popular Models
MiniMax M3$60.00/M
MiniMax M3$60.00/M
Buyer's guide
Choosing an API Provider
Key factors for selecting an inference provider
Pricing models
Providers charge per-token (input/output priced separately), per-request, or offer committed use discounts. For high-volume apps, $0.50/M token differences translate to thousands in monthly savings.
Latency & throughput
First-token latency matters for interactive apps; total generation time for batch processing. Throughput (tokens/sec) is critical for real-time applications and agent workflows.
Model selection
First-party providers (OpenAI, Anthropic) offer latest models first. Third-party providers (Together, Fireworks, Groq) often provide same quality at lower cost plus open-source alternatives.
Reliability & support
Uptime, rate limits, and SLAs vary significantly. For production workloads, consider multi-provider strategies with automatic failover. Check our provider rankings.
FAQ
Frequently Asked Questions
Common questions about LLM updates, version releases, and API changes
What are the latest LLM version updates?
⌄
LLM Stats tracks all major language model version releases in real-time. This includes updates from OpenAI (GPT series, o-series), Anthropic (Claude), Google (Gemini), Meta (Llama), Mistral, DeepSeek, and other providers. Our timeline shows release dates, new capabilities, benchmark improvements, and feature additions for each version.
How do I track API changes and pricing updates?
⌄
Our API Provider Updates section tracks pricing changes, new features, rate limit updates, and API endpoint modifications for all major providers including OpenAI, Anthropic, Google Cloud AI, Azure OpenAI, Together AI, Fireworks, Groq, and more. We also monitor throughput and latency changes.
How do I compare different versions of an LLM?
⌄
Use our Model Comparison tool to see side-by-side benchmark performance, capability differences, pricing changes, and context window improvements between model versions. You can compare versions like GPT-4 vs GPT-4 Turbo vs GPT-4o, or track improvements across Claude 3 vs Claude 3.5 releases.
Which organizations release the most LLM updates?
⌄
Our Active Organizations section shows the most active AI labs by recent releases. Major contributors include OpenAI, Anthropic, Google DeepMind, Meta AI, Mistral AI, xAI, DeepSeek, Alibaba (Qwen), and many others. Filter by organization to see their complete release history and model families.
How often is this page updated?
⌄
LLM Stats updates this page hourly to capture new model releases and API changes as they happen. Our team also manually verifies major announcements from provider blogs, press releases, and official documentation. For real-time updates, follow us on Twitter.
More
Explore More
Dive deeper into LLM data, benchmarks, and comparisons