StepFun releases Step 3.7 Flash
On 29 May 2026 StepFun released Step 3.7 Flash under the tagline "The new frontier is agent efficiency". The model adds image understanding to Step 3.5 Flash and introduces an Advisor Mode for agent tasks.
Total 1,098
Anthropic releases Claude Opus 4.8
On 28 May 2026 Anthropic released Claude Opus 4.8, a modest improvement on Opus 4.7 at the same price. Fast mode became three times cheaper, and Claude Code gained dynamic workflows that run hundreds of parallel subagents.
Mistral merges Le Chat into Vibe, a unified agent for work and code
On 28 May 2026 Mistral merged Le Chat into Vibe: one agent and one licence spanning work and code, adding Work Mode and sandboxed Code Mode, four tiers, and the retirement of the old Chat mode.
Mistral AI agrees to acquire physics-AI company Emmi AI
On 23 May 2026 Mistral AI said it had agreed to acquire Emmi AI, an Austrian physics-AI company whose large engineering models simulate power grids, injection moulding and car safety tests; more than 30 staff, including co-founder Johannes Brandstetter, are joining.
OpenBMB publishes MiniCPM5-1B
OpenBMB published MiniCPM5-1B on 21 May 2026, the first model in the MiniCPM5 series: a dense 1B Transformer for on-device use, with Apache-2.0 weights and UltraData training data.
Stability AI releases Stable Audio 3.0
In May 2026 Stability AI released Stable Audio 3.0 in four sizes, three of them open-weight. The two larger models generate up to 6 minutes 20 seconds of audio and support inpainting and continuation.
Cohere releases Command A+ as an Apache 2.0 open model
On 20 May 2026 Cohere released Command A+ as an Apache 2.0 open model with 218 billion total parameters. It was aimed at sovereign and critical-infrastructure users.
Google introduces Gemini 3.5 Flash at I/O 2026
At I/O on 19 May 2026 Google announced the Gemini 3.5 family, led by Gemini 3.5 Flash, made it the default in the Gemini app and Search's AI Mode, and introduced Gemini Spark and Antigravity 2.0.
Google launches Gemini Omni for conversational video creation
On 19 May 2026 Google added Gemini Omni to the Gemini app, letting subscribers create and edit video through conversational prompts from text, photos and clips, with SynthID watermarks and C2PA Content Credentials attached to the output.
Google introduces Gemini Spark, a 24/7 personal agent
At Google I/O on 19 May 2026, Google introduced Gemini Spark, a personal AI agent built on Gemini 3.5 Flash that runs around the clock and acts on the user's behalf under their direction. It launched first with trusted testers and US Google AI Ultra subscribers.
Jury rejects Elon Musk's lawsuit against OpenAI
A jury in Oakland ruled for OpenAI, Sam Altman, Greg Brockman and Microsoft on 18 May 2026, finding that Elon Musk's claims were filed too late under the three-year statute of limitations. Musk says he will appeal.
Isomorphic Labs raises a $2.1 billion Series B
On 12 May 2026 Isomorphic Labs announced a $2.1 billion Series B round led by Thrive Capital, money the company says will scale its business globally and push its drug candidate pipeline forward.
Baidu releases ERNIE 5.1
Baidu released ERNIE 5.1 on 9 May 2026, a model derived from ERNIE 5.0's pre-training with about a third of its total parameters and much lower training cost, after a preview build topped Chinese models on LMArena.
StepFun completes a US$2.5 billion round ahead of a Hong Kong IPO
In May 2026 The Standard reported that StepFun had completed a funding round of about US$2.5 billion ahead of a planned Hong Kong listing, after converting into a joint-stock company in April.
ElevenLabs passes $500M ARR and adds BlackRock, NVIDIA and others as investors
On 5 May 2026 ElevenLabs said it had crossed $500 million in annual recurring revenue and closed a third tranche of its Series D, with BlackRock, NVIDIA's NVentures, Santander and others investing.
Kunlunxin starts STAR Market IPO guidance
In May 2026 Baidu's chip unit Kunlunxin began IPO guidance for China's STAR Market while keeping its Hong Kong plan, as revenue grows and losses narrow toward break-even.
Anthropic raises $65 billion
In May 2026 Anthropic raised $65 billion at a $965 billion valuation, a Series H round that mixed new capital with $15 billion of earlier hyperscaler commitments, including $5 billion from Amazon.
Poolside releases its first two Laguna coding models
On 28 April 2026 Poolside released Laguna M.1 and Laguna XS.2, its first two coding models, together with pool, the agent runtime its own reinforcement learning runs on, as a research preview.
Mistral AI releases open-weight Mistral Medium 3.5
Mistral Medium 3.5 is a dense 128B model released by Mistral AI on 28 April 2026, with a 256K context, text and image input, and open weights under a modified MIT licence.
SenseTime open-sources the SenseNova U1 unified multimodal model
On 27 April 2026 SenseTime released the initial weights and inference code of SenseNova U1, an open-source unified multimodal model under the Apache License 2.0.
Microsoft and OpenAI amend their deal, ending cloud exclusivity
On 27 April 2026 Microsoft and OpenAI amended their partnership so OpenAI can sell on any cloud, ending Azure's exclusivity while Microsoft keeps a non-exclusive licence to its models.