0
SIGNAL + NOISE / INDEX 013

Model Signal

Model releases, translated into operating decisions.

A recurring Signal + Noise series on frontier, open, multimodal, and specialist models — what changed, where each model fits, what breaks, and what teams should do next.

TRACKED059
LATESTQwen3.8-Max
UPDATEDJuly 17, 2026
LATEST MODEL SIGNAL·July 17, 2026·NEW

Qwen3.8-Max

Alibaba's 2.4-trillion-parameter flagship large language model, released in preview to compete directly with leading frontier systems.

ALIBABAGENERAL
JULY 2026 · 8 OF 59 TRACKED
Alibaba·July 17, 2026·NEW

Qwen3.8-Max

Alibaba's 2.4-trillion-parameter flagship large language model, released in preview to compete directly with leading frontier systems.

General
Moonshot AI·July 16, 2026·NEW

Kimi K3

Moonshot AI has released Kimi K3, a 2.8-trillion parameter open-weight model that achieves frontier-level coding capabilities and outperformed Claude Fable 5 on the Frontend Code Arena.

General·1000000
OpenAI·July 9, 2026·NEW

ChatGPT Work

ChatGPT Work is an OS-level AI agent by OpenAI that integrates across Mac and Windows applications to automate research and execute multi-step document creation tasks for business workflows.

General·Balanced/Terra for Free/Go, Sol/Terra/Luna for Plus/Pro/Business/Enterprise
Meta·July 9, 2026·NEW

Muse Spark 1.1

Meta's Muse Spark 1.1 is an API-accessible, agentic coding model featuring a 1-million-token context window and optimized for automated tool and computer use.

Code
OpenAI·July 9, 2026·NEW

GPT-5.6 Sol

GPT-5.6 Sol is OpenAI's flagship frontier model, offering top-tier capabilities in complex coding, scientific reasoning, and cybersecurity.

General
OpenAI·July 9, 2026·NEW

GPT-5.6

OpenAI's GPT-5.6 is a major frontier model family featuring three tiers—Sol, Terra, and Luna—designed to deliver improved intelligence and cost efficiency across enterprise workloads.

General
OpenAI·July 8, 2026·NEW

GPT-Live

OpenAI's GPT-Live is a family of full-duplex voice models designed for low-latency, real-time human-AI interaction.

Multimodal
Base44·July 1, 2026

Base 1

Base 1 is a proprietary code generation model trained by Base44 to natively power its web application development and 'vibe coding' workflows.

Code
JUNE 2026 · 11 OF 59 TRACKED
Google·June 30, 2026

TabFM

Google's TabFM is a zero-shot foundation model capable of performing classification and regression on tabular data without requiring per-dataset training.

General·N/A
Google·June 30, 2026

Nano Banana 2 Lite

A fast, cost-efficient text-to-image model from Google designed to generate outputs in approximately four seconds.

Image
Anthropic·June 30, 2026

Claude Sonnet 5

Anthropic's Claude Sonnet 5 introduces advanced agentic capabilities and a one-million token context window at a lower cost profile than its Opus counterparts.

General·1M tokens
Liquid AI·June 25, 2026

LFM2.5-230M

Liquid AI has released a 230-million parameter open-weights model optimized for local data extraction and on-device agentic workflows.

General
Sakana AI·June 22, 2026

Sakana Fugu

Sakana Fugu is an orchestration model that functions as a multi-agent system by dynamically routing tasks across a swappable pool of underlying large language models.

General
WeiboAI·June 17, 2026

VibeThinker-3B

VibeThinker-3B is a 3-billion-parameter open-weights reasoning model developed by WeiboAI.

Reasoning
Z.ai·June 13, 2026

GLM-5.2

GLM-5.2 is a 753-billion parameter open-weights large language model from Z.ai that is specifically optimized for long-horizon coding tasks.

Code·1000000
Avataar AI·June 11, 2026

Varya

An open-weight video generation model optimized for Indian cultural contexts and highly cost-efficient inference.

Multimodal
Anthropic·June 9, 2026

Claude Mythos 5

Anthropic's frontier Claude Mythos 5 model was launched and subsequently disabled worldwide following a US government security directive.

Reasoning
Anthropic·June 9, 2026

Claude Fable 5

Anthropic's most powerful publicly released model — its first Mythos-class model available to enterprise and paid users. Leads on SWE-bench, knowledge work, and scientific research, scoring 10%+ above Opus 4.8 on key benchmarks.

Reasoning·1M
NVIDIA·June 4, 2026

Nemotron-3-Ultra-550B-A55B-NVFP4

A frontier-scale 550-billion parameter hybrid model from NVIDIA featuring a 1-million token context window and NVFP4 precision.

General·512288
MAY 2026 · 3 OF 59 TRACKED
APRIL 2026 · 9 OF 59 TRACKED
DeepSeek·April 24, 2026

DeepSeek V4 Flash

Reasoning-optimized flash variant of DeepSeek V4 — neck-and-neck with Kimi K2.6 on coding benchmarks at significantly lower latency.

Reasoning·128K
OpenAI·April 23, 2026

GPT-5.5

OpenAI's latest flagship with noticeably stronger reasoning and autonomy than GPT-5.4. Available in standard and Pro variants. 1M context window, $5/$30 per 1M tokens for standard.

Reasoning·1M
Qwen·April 21, 2026

Qwen3.6-27B

A 27.8-billion parameter dense multimodal model optimized for local deployment with strong coding capabilities and a 256K context window.

Multimodal·256K
Qwen·April 16, 2026

Qwen3.6-35B-A3B

A mid-sized, open-weight multimodal mixture-of-experts model from Qwen featuring 35 billion total and 3 billion active parameters.

Multimodal·262144
Z.ai·April 7, 2026

GLM-5.1

Open-weights model from Z.ai (Zhipu AI) with strong web-dev coding performance. Ranks above Kimi K2.6 on Code Arena WebDev leaderboard (1,534 Elo). Competitive on multilingual tasks.

Code·200K
Google·April 3, 2026

gemma-4-26B-A4B-it

A 26-billion-parameter, instruction-tuned Mixture-of-Experts model from Google supporting multimodal inputs and a 256K-token context window.

Multimodal·256K
Google·April 2, 2026

gemma-4-31B-it

An instruction-tuned 31-billion parameter open-weights model from Google's Gemma 4 family, featuring native multimodal support and a 256K context window.

Multimodal·256K
Moonshot AI·April 1, 2026

Kimi K2.6

1 trillion-parameter vision-language model from Moonshot AI. Highest-ranked open-weights model on Artificial Analysis leaderboard (score 54). Designed for long-horizon agentic coding with plan-write-test-debug loops lasting days.

Code·128K
DeepSeek·April 1, 2026

DeepSeek V4

DeepSeek's latest open-source flagship. Rivals closed frontier models on coding and reasoning benchmarks while maintaining extremely low inference cost. MoE architecture.

Code·128K
MARCH 2026 · 4 OF 59 TRACKED
FEBRUARY 2026 · 4 OF 59 TRACKED
DECEMBER 2025 · 1 OF 59 TRACKED
NOVEMBER 2025 · 1 OF 59 TRACKED
OCTOBER 2025 · 1 OF 59 TRACKED
SEPTEMBER 2025 · 1 OF 59 TRACKED
AUGUST 2025 · 2 OF 59 TRACKED
JULY 2025 · 3 OF 59 TRACKED
MAY 2025 · 1 OF 59 TRACKED
APRIL 2025 · 5 OF 59 TRACKED
FEBRUARY 2025 · 2 OF 59 TRACKED
DATE PENDING · 3 OF 59 TRACKED
About Model Signal

Model Signal tracks AI model releases across every provider. Each model has a compounding Model Signal report — a synthesized brief that auto-updates as the Signal + Noise wire mentions it. It covers releases through an operator lens: who should care, where each model fits, what changes, and what does not.