AI model news, week of 28 September 2026
In the week of 28 September 2026, OpenAI, Anthropic, and xAI published 7 updates that matter for developers: new models (2), deprecations and retirements (2), and announcements (3).
New models
- GPT-6.1 Sol released in the API for complex coding and professional work
OpenAI released GPT-6.1 Sol, model gpt-6.1-sol, at $2.00 input and $10.00 output per 1M tokens, below the price of GPT-6 Astra.
- Anthropic launches Claude Sonnet 5.5 (claude-sonnet-5-5)
Claude Sonnet 5.5 is now available on the Claude API and other platforms, with a 1M token context window and $2 / MTok input pricing.
Deprecations and retirements
- xAI deprecates grok-voice-transcribe-1.0
grok-voice-transcribe-1.0 is deprecated and will reach end of life on October 2, 2026, with requests routed to grok-voice-transcribe-2.0.
- Claude Sonnet 4.5 is deprecated, retiring on November 30, 2026
Anthropic deprecated Claude Sonnet 4.5 in the Claude API, with retirement set for November 30, 2026, and recommends Claude Sonnet 5.5.
Announcements
- OpenAI adds computer use to the Agents API
OpenAI added computer use to the Agents API, allowing agents to complete tasks in a hosted browser.
- OpenAI adds Ultrafast mode for GPT-6 Astra
OpenAI added Ultrafast mode for GPT-6 Astra in the Responses API to lower output token latency.
- OpenAI simplifies API usage tiers from five to three
OpenAI simplified its API usage tiers to three tiers: Build, Launch, and Grow.
Models from this week on Zabio
| Model | Input, $/M | Output, $/M | Cached, $/M | Context window | 1,000 calls | Writing code, hard |
|---|---|---|---|---|---|---|
gpt-6.1-solOpenAI | 2.00 | 10.00 | 0.10 | 400,000 | $9.00 | 63.4 |
gpt-6-astraOpenAI | 10.00 | 50.00 | 1.00 | 400,000 | $45.00 | 67.8 |
claude-sonnet-5-5Anthropic | 2.00 | 10.00 | 0.10 | 1,000,000 | $9.00 | 80.6 |
claude-sonnet-5Anthropic | 2.00 | 10.00 | 0.20 | 1,000,000 | $9.00 | 63.3 |
Prices are the vendor list prices in USD per million tokens. "1,000 calls" is 1,000 requests of 2,000 input tokens and 500 output tokens each, without caching. Zabio bills the list price plus its service fee: see pricing.
Scores: percent correct for that kind of job, aggregated from public results by Epoch AI, LMArena, and LiveBench, read on 16 November 2026. Epoch AI: CC BY 4.0, LMArena: CC BY 4.0, LiveBench: Leaderboard results used with attribution; the 2.0 release states no terms of its own (read 2026-11-06); the first version's Mathematics, Language, IF results: Code Apache 2.0 and MIT; leaderboard results used with attribution