codingsalt

LLM API Price Tracker

Most pricing pages show today's rate. This one shows when it changed: a dated timeline of every LLM API price event we've covered, each linked to the article that documented it the day it happened. Last updated 2026-09-03.

Want today's rates against your own token counts? Use the LLM API cost calculator. Building on this data? It's free as JSON at /api/model-prices.json.

Scheduled changes ahead

Already announced by the vendor, not yet in effect — budget for these now.

  • Scheduled +100%

    Gemini 3.8 Flash (Google) — $0.75/$3.75 $1.5/$7.5 per 1M tokens

    Gemini 3.8 Flash's introductory rate expires December 31, 2026; from January 1, 2027 Google charges $1.50/$7.50 — a 100% increase on both sides.

    Read the coverage

The timeline

Newest first. USD per million tokens, cache-miss list rates.

  • Launch

    Gemini 3.8 Flash (Google) — $0.75/$3.75 per 1M tokens

    Gemini 3.8 Flash launched at an introductory $0.75/$3.75 — matching 3.7 Flash's introductory rate — valid through December 31, 2026.

    Read the coverage

  • Price increase +50%

    Claude Sonnet 5 (Anthropic) — $2/$10 $3/$15 per 1M tokens

    Claude Sonnet 5's introductory pricing ended: the standard $3/$15 rate took effect, a 50% increase over the $2/$10 launch rate — a scheduled reversion published at launch, not a surprise hike.

    Read the coverage

  • Price held

    DeepSeek V4 Flash (DeepSeek) — $0.14/$0.28 per 1M tokens

    DeepSeek V4 Flash exited preview as build 0731 with pricing held at $0.14/$0.28 ($0.0028 input on cache hits); a 2x peak-hours multiplier is documented but has no effective date.

    Read the coverage

  • Price cut -80%

    GPT-5.6 Luna (OpenAI) — $1/$6 $0.2/$1.2 per 1M tokens

    OpenAI cut GPT-5.6 Luna by 80%, from $1.00/$6.00 to $0.20/$1.20, three weeks after the GPT-5.6 family reached general availability.

    Read the coverage

  • Price cut -20%

    GPT-5.6 Terra (OpenAI) — $2.5/$15 $2/$12 per 1M tokens

    OpenAI cut GPT-5.6 Terra by 20%, from $2.50/$15.00 to $2.00/$12.00, in the same July 30 announcement as the Luna cut; Sol stayed at $5/$30.

    Read the coverage

  • Launch

    Claude Opus 5 (Anthropic) — $5/$25 per 1M tokens

    Claude Opus 5 launched at Opus 4.8's unchanged $5/$25 rate — a model-ID swap with no pricing change for existing Opus users.

    Read the coverage

  • Launch

    Claude Sonnet 5 (Anthropic) — $2/$10 per 1M tokens

    Claude Sonnet 5 launched at an introductory $2/$10, with the standard $3/$15 rate documented from day one as taking effect September 1, 2026.

    Read the coverage

Frequently asked questions

When did GPT-5.6 Luna's price drop?
On July 30, 2026, OpenAI cut GPT-5.6 Luna by 80%, from $1.00/$6.00 to $0.20/$1.20 per million input/output tokens. GPT-5.6 Terra was cut 20% in the same announcement, from $2.50/$15.00 to $2.00/$12.00.
When did Claude Sonnet 5 get more expensive?
On September 1, 2026, Claude Sonnet 5's introductory $2/$10 pricing ended and Anthropic's standard $3/$15 rate took effect — a 50% increase that was documented as scheduled when the model launched on June 30, 2026.
Which LLM price changes are already scheduled?
Gemini 3.8 Flash's introductory $0.75/$3.75 rate expires December 31, 2026; from January 1, 2027 Google charges $1.50/$7.50 — a 100% increase. DeepSeek has also documented a 2x peak-hours multiplier for V4 Flash without announcing an effective date.
Where does this price history come from?
Every event on this page is documented in a dated CodingSalt article published when the change happened, sourced from the vendor's official pricing page or announcement. Nothing is listed without a linked, published source.

The timeline grows as prices move — we publish changes the day they happen. Follow the RSS feed to catch the next one.