AI model news, week of 5 October 2026
In the week of 5 October 2026, Anthropic, Google, OpenAI, Mistral AI, and Together AI published 10 updates that matter for developers: new models (5), price changes (1), deprecations and retirements (1), and announcements (3).
New models
- chat-latest now points to the latest model used in ChatGPT
OpenAI updated the chat-latest alias, which now points to the latest Instant model used in ChatGPT. The price is $5.00 input and $30.00 output per 1M tokens.
- Anthropic releases Claude Haiku 5.5 in the Claude API
Claude Haiku 5.5 is now available in the Claude API with a 1M token context window and up to 128k output tokens.
- OpenAI releases Decisions API beta with gpt-6-luna for typed answers
The Decisions API is in public beta and returns typed answers from text and images, with gpt-6-luna as the only model currently available.
- Google releases Gemini Nano Banana 2.1 and deprecates older image model
Google released Gemini Nano Banana 2.1 for image generation and announced that gemini-3.1-flash-image will shut down on October 29, 2026.
- Mistral launches Mistral Large 4 public preview
Mistral launched a public preview API for Mistral Large 4, priced at $1.36 per million input tokens and $4.18 per million output tokens.
Price changes
- Lower prompt cache read price on Claude Sonnet 5.5
Anthropic cut the price of prompt cache reads on Claude Sonnet 5.5 from $0.20 USD to $0.10 USD per million tokens.
Deprecations and retirements
- Google deprecates gemini-3.1-flash-image
Google has deprecated the gemini-3.1-flash-image model and recommends migrating to gemini-nano-banana-2.1.
Announcements
- Anthropic adds thinking.types.disabled parameter to Models API
Anthropic updated its Models API endpoints to report whether each model accepts disabling thinking mode.
- Anthropic expands the Cyber Verification Program
Anthropic launched an expanded Cyber Verification Program with three access tiers for security teams.
- Together AI launches Together Link for coding agents
Together AI introduced Together Link, a tool connecting coding agents to models like GLM 5.3 and Kimi K3 using an API key.
Models from this week on Zabio
| Model | Input, $/M | Output, $/M | Cached, $/M | Context window | 1,000 calls | Analysis and reasoning, hard |
|---|---|---|---|---|---|---|
chat-latestOpenAI | 5.00 | 30.00 | 0.50 | 400,000 | $25.00 | no data |
gpt-6-astraOpenAI | 10.00 | 50.00 | 1.00 | 400,000 | $45.00 | 93.0 |
claude-sonnet-5-5Anthropic | 2.00 | 10.00 | 0.10 | 1,000,000 | $9.00 | 91.7 |
claude-haiku-5-5Anthropic | 0.10 | 0.50 | 0.01 | 1,000,000 | $0.45 | 85.5 |
claude-haiku-4-5Anthropic | 1.00 | 5.00 | 0.10 | 200,000 | $4.50 | 60.5 |
gpt-6-lunaOpenAI | 0.10 | 0.50 | 0.01 | 400,000 | $0.45 | 73.6 |
gemini-nano-banana-2.1Google | 1.50 | 7.50 | no data | 0 | $6.75 | no data |
claude-opus-5-5Anthropic | 4.00 | 20.00 | 0.20 | 1,000,000 | $18.00 | 90.6 |
deepseek-v4-pro-0813DeepSeek | 1.32 | 3.96 | 0.264 | 131,072 | $4.62 | 91.7 |
kimi-k3Moonshot AI | 3.00 | 15.00 | 0.30 | 1,048,576 | $13.50 | 85.1 |
glm-5.3Z.ai | 1.40 | 4.40 | 0.26 | 1,048,575 | $5.00 | 90.9 |
glm-5.2Z.ai | 1.40 | 4.40 | 0.26 | 1,048,575 | $5.00 | 71.2 |
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