Mikä tekoälyssä muuttui

Proof of work — the maintenance ledger

Every entry below is computed from a dated record — a committed price-refresh file, an approved radar update, or a source-cited model release. Nothing here is typed in by hand.

1. 2026-10-07Claude Haiku 5.5 shipped — Anthropic's 7 Oct 2026 small model, on the Claude API as claude-haiku-5-5 and on AWS, Google Cloud and Azure, with a 1M-token context window, 128K output and, for the first time on a Haiku, effort settings (default medium) and adaptive thinking. It is the first Claude priced by prompt length: a prompt over 100K tokens pays 5x the base rate. Anthropic reports 72.4% on the OSWorld 2.1 offline subset against Haiku 4.5's 15.7%; that is Anthropic's own figure.
2. 2026-10-06Mistral Large 4 shipped — Mistral's 6 Oct 2026 open-weight hybrid instruct-and-reasoning model: a mixture-of-experts with 1 trillion total and 52 billion active parameters and multimodal input. At launch it is a public API preview; Mistral says the weights "drop end of this month". All benchmark figures here are Mistral's own.
3. 2026-10-01Prices re-verified for 5 tools — 22 plans re-checked against the new prices
4. 2026-09-30Gemini 4 Argon shipped — Google's 30 Sep 2026 flagship, built for long multi-step work in software engineering, legal and finance research, and cyber defence, with a 1M-token output limit (up from 64K). Google also reports 51.3% on Zapier's AutomationBench (ranked #1), 91.7% on LVBench long-video understanding and 68% on CWE-bench v1 vulnerability fixing (tied first); those are Google's own figures. At launch it is only rolling out to vetted cyber defenders through Google's Fairwind Program while it goes through the US government's voluntary pre-release review; Google says paid API customers and Google AI Ultra subscribers come first when it opens, with no date.
5. 2026-09-29GPT-6.1 Sol shipped — OpenAI's 29 Sep 2026 DevDay upgrade to GPT-6 Sol, on the API as gpt-6.1-sol with a 1,050,000-token window, 128,000-token output and an April 2026 knowledge cutoff, and in ChatGPT Work and Codex for Plus, Pro, Business, Enterprise and Edu (not yet in regular ChatGPT chat). OpenAI reports it within 2.1 points of GPT-6 Astra on OSWorld 2.0 computer use at about one-seventh the cost per task, 2.2 points above Claude Opus 5.5 on AutomationBench at about a third of the cost, and $5.47 per Terminal-Bench Science task against $23.21 for Opus 5.5; those are OpenAI's own figures, and Astra still scores highest on the hardest science tasks (68.1%). A faster Ultrafast tier is promised "in the coming days".
6. 2026-09-28Claude Sonnet 5.5 shipped — Anthropic's 28 Sep 2026 mid-tier model, on the Claude API as claude-sonnet-5-5 with a 1M-token context window and 128K output, at the same $2 / $10 as Sonnet 5. Anthropic reports it 30%+ faster and up to 30% cheaper per task than Sonnet 5, and within 2 points of Opus 5.5 on GDPval-AA knowledge work (1844 against 1846) but behind it on CursorBench 4.0 (55.5% against 57.8%); those are Anthropic's own figures. Anthropic says Opus 5.5 "remains clearly stronger at complex, open-ended work requiring sustained judgment". Claude Haiku 5.5 followed on 7 Oct 2026.
7. 2026-09-25Prices re-verified for 14 tools — 23 plans re-checked against the new prices
8. 2026-09-25MiniMax price rose: $20/mo → $22/mo — The cheapest paid plan for MiniMax moved from $20/mo to $22/mo (+10%), verified on the vendor's pricing page on 2026-09-25.

Poistuvat mallit

• DeepSeek V4 Flash (0731) poistuu 2026-09-10 — vaihda malliin DeepSeek V4.1 Flash.
• DeepSeek V4 poistuu 2026-09-10 — vaihda malliin DeepSeek V4.1 Flash.

Koko mallien aikajana