AI updates
AI models, chat apps, and voice and video tools. Price changes, launches, policy changes and shutdowns, each dated and taken from the company's own pages. Prices in an update are as of its date; the comparison pages always show today's.
October 2026
Kling 4.0 is out for everyone, with 30-second clips and 4K output
Kling AI made Kling 4.0 available to all users on October 15, 2026. It makes clips of 3 to 30 seconds at up to 4K, with stereo sound and up to 10 keyframes.
OpenAI will remove GPT-4o and GPT-4.1 from its API on January 14, 2027
On October 14, 2026, OpenAI deprecated GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini. All four leave the API on January 14, 2027; OpenAI points users to GPT-6 Luna.
Google's Gemini 4 Argon is out at $2/$10; the API price doubles December 1
Gemini 4 Argon became generally available on October 13, 2026 at $2 input and $10 output per 1M tokens through November 30. From December 1 it costs $4 and $20.
Meta's Muse Wonder: API price and specs
Muse Wonder costs $4 per million input tokens and $20 per million output tokens on Meta's API, with a 1,048,576-token context window. Meta still labels it a preview.
Google retires Gemini 3.7 Flash and 3.5 Flash; API calls go to newer models
Since October 8, 2026, Gemini API requests to gemini-3.7-flash go to Gemini 3.8 Flash and requests to gemini-3.5-flash go to Gemini 3.6 Flash, without a code change.
Anthropic's Claude Haiku 5.5 costs $0.10/$0.50, a tenth of Haiku 4.5
Claude Haiku 5.5 launched on October 7, 2026 at $0.10 input and $0.50 output per 1M tokens for prompts up to 100,000 tokens. Haiku 4.5 costs $1 and $5.
Anthropic halves Claude Sonnet 5.5's cache read price to $0.10
From October 7, 2026, prompt cache reads on Claude Sonnet 5.5 cost $0.10 per 1M tokens instead of $0.20. Input, output and cache write prices are unchanged.
ChatGPT moves every plan to GPT-6: Sol for paid plans, Luna for Free and Go
From October 7, 2026, GPT-6 Sol replaces GPT-5.6 Sol in ChatGPT for paid plans. Free and Go users get GPT-6 Luna from October 8.
Claude Max and Team plans now include $100 to $500 a month in API credits
Since October 7, 2026, Claude Max 5x includes $100 and Max 20x $200 of monthly API credits. Team plans get $20 or $100 per seat, pooled up to $500. Credits don't roll over.
Mistral Large 4 opens as a public API preview, with open weights due in October
Mistral released Mistral Large 4 as a public preview on October 6, 2026: a 1-trillion-parameter model listed at $1.36 input and $4.18 output per 1M tokens, now on sale at half that.
OpenAI will remove tts-1, tts-1-hd and gpt-4o-mini-tts on January 6, 2027
On October 1, 2026, OpenAI deprecated its four text-to-speech models. They leave the API on January 6, 2027, and OpenAI points users to gpt-realtime-2.1-mini.
September 2026
Anthropic will retire Claude Sonnet 4.5 on November 30, 2026
On September 30, 2026, Anthropic deprecated Claude Sonnet 4.5. Requests to it on the Claude API will fail after November 30, 2026; the named replacement is Sonnet 5.5.
ElevenLabs launches Eleven v4 and v4 Turbo, with 90+ languages
ElevenLabs released Eleven v4 and the low-latency Eleven v4 Turbo on September 28, 2026. Both cover 90+ languages, and Instant Voice Clones now need 10 seconds of audio.
OpenAI shut down the Sora 2 video API on September 24, 2026
OpenAI removed the Videos API and all Sora 2 models from its API on September 24, 2026, with no replacement model. The Sora app and website closed on April 26, 2026.
Anthropic's Claude Opus 5.5 launches at $4/$20, 20% below Opus 5
Claude Opus 5.5 arrived on September 22, 2026 at $4 input and $20 output per 1M tokens, down from $5 and $25 for Opus 5, with a 1M-token context window.
Google's Gemini 3.8 Flash TTS is out; its API price doubles in 2027
Gemini 3.8 Flash TTS and Flash-Lite TTS became generally available on September 22, 2026. Audio output costs $9 and $6 per 1M tokens through December 31, then doubles.
OpenAI's GPT-6 Sol and Luna arrive at half the price of GPT-5.6
On September 22, 2026, OpenAI released GPT-6 Sol at $2/$10 and GPT-6 Luna at $0.10/$0.50 per 1M tokens. GPT-6.1 Sol followed on September 29 at the same base price.
Grok 4.7 launches at Grok 4.6's price: $2 input, $6 output per 1M tokens
Grok 4.7 was released on September 21, 2026. In the API it costs $2 input and $6 output per 1M tokens below 200K-token prompts and has a 500K context window.
Descript now bills jump-cut smoothing by the second
Since September 17, 2026, Descript charges 0.5 AI credits per second of regenerated audio and 1.0 per second of video to smooth jump cuts, instead of a flat charge per cut.
Pika relaunches as a creative platform with 25+ apps for video, image and sound
On September 17, 2026, Pika replaced its app with the New Pika: 25+ apps, every model on every plan, and clips up to 30 seconds. Existing plans carry over at the same price.
Suno launches v6: v6 and v6-wild for paid plans, v6-mini for everyone
Suno released its v6 music models on September 9, 2026, made with Warner Music Group, BMG and Believe. Pro and Premier get v6 and v6-wild; free users get v6-mini.
Runway now runs inside Adobe Premiere Pro and After Effects
On September 8, 2026, Runway released a free plugin panel for Premiere Pro and After Effects on macOS and Windows. Generations use the credits on your paid Runway plan.
Runway adds a Team plan at $69 per seat with pooled credits
Runway launched a self-serve Team plan on September 4, 2026: $69 per seat a month, or $55 billed annually, for 2 to 9 seats. Each seat adds 6,900 credits to a shared pool.
Google's Lyria 3.5 music model is on the Gemini API at $0.08 per song
Google made Lyria 3.5 generally available on the Gemini API on September 3, 2026. It makes full songs with vocals and lyrics, a couple of minutes long, for $0.08 per song.
OpenAI releases GPT-6 Astra at $10 input and $50 output per 1M tokens
GPT-6 Astra reached OpenAI's API on September 3, 2026. It costs $10 input and $50 output per 1M tokens and has a 1,050,000-token context window.
Google's Gemini 3.8 Flash: $0.75/$3.75 now, double from January 2027
Gemini 3.8 Flash became generally available on September 2, 2026. It costs $0.75 input and $3.75 output per 1M tokens through 2026, then $1.50 and $7.50 from January 1, 2027.
August 2026
Gemini Omni Flash video model goes GA at about $0.10 per second
On August 27, 2026, Google released gemini-omni-1.1-flash, the GA version of its Omni Flash video model, on the Gemini API. 720p video costs about $0.10 a second.
DeepSeek adds peak and off-peak API prices; off-peak costs half
Since August 16, 2026, DeepSeek's API charges full price in two weekday windows (01:00-04:00 and 06:00-10:00 UTC) and half price at all other times, including US business hours.
Alibaba releases its Qwen3.8 flagship as open weights
Alibaba added Qwen3.8-2.4T-A95B, the open-weight version of its Qwen3.8 flagship, to Model Studio on August 12, 2026. Hosted, it costs $2 input and $6 output per 1M tokens.