AI Model Pricing Changelog
Every major pricing change across Claude, GPT, Gemini, DeepSeek and more — tracked since 2023.
Last updated: 15 Jun 2026
32
Price changes tracked
−70%
Biggest single cut — Mistral Small 3.2
−92%
Flagship input price since 2023
7
Providers tracked
Flagship input price over time
$/MTok input — Anthropic, OpenAI, Google and DeepSeek flagship models
Provider
Type
Year
| Date | Provider | Model | Change | Before | After | Delta | Note |
|---|---|---|---|---|---|---|---|
| 15 Jun 2026 | Anthropic | Claude Subscriptions | PRICE INCREASE | — | — | — | Anthropic splits subscription quota into two pools — automated/agentic usage decoupled from interactive access. Previously all shared one quota. Effective June 15, 2026. |
| 19 May 2026 | Google AI Ultra | PRICE CUT | — | — | — | Google I/O: AI Ultra subscription cut 20% from $250 to $200/month. New entry-level AI Ultra tier at $100/month. | |
| 16 Apr 2026 | Anthropic | Claude Opus 4.7 | LAUNCH | — | $5.00 / $25.00 | — | Claude Opus 4.7 — new tokenizer, up to 35% more tokens for same input text vs Opus 4.6 |
| Apr 2026 | Anthropic | Claude Sonnet 4.6 | LAUNCH | — | $3.00 / $15.00 | — | Claude Sonnet 4.6 — Sonnet line holds at $3/M for third consecutive year |
| Mar 2026 | xAI | Grok 4.3 | LAUNCH | — | $1.25 / $2.50 | — | Grok 4.3 — xAI competitive mid-tier entry |
| Feb 2026 | Gemini 3.1 Pro | LAUNCH | — | $2.00 / $12.00 | — | Gemini 3.1 Pro — updated flagship | |
| Jan 2026 | Gemini 3 Flash | LAUNCH | — | $0.50 / $3.00 | — | Gemini 3 Flash — efficient tier upgrade | |
| Jan 2026 | OpenAI | GPT-5.4 | LAUNCH | — | $2.50 / $15.00 | — | GPT-5.4 — 1.05M context, flagship repositioned upward from GPT-5 |
| Nov 2025 | Anthropic | Claude Haiku 4.5 | LAUNCH | — | $1.00 / $5.00 | — | Claude Haiku 4.5 — updated budget tier |
| Sept 2025 | OpenAI | GPT-5.2 | LAUNCH | — | $1.75 / $14.00 | — | GPT-5.2 — capability upgrade, slight price increase over GPT-5 |
| Jun 2025 | Gemini 2.5 Pro | LAUNCH | — | $1.25 / $10.00 | — | Gemini 2.5 Pro — 2M context window at $1.25/M | |
| Jun 2025 | Mistral | Mistral Small 3.2 | PRICE CUT | $0.20 / $0.60 | $0.06 / $0.18 | -70% | Mistral Small 3.2 — 70% price cut, one of the cheapest capable models available |
| Apr 2025 | Anthropic | Claude Opus 4.5 | PRICE CUT | $15.00 / $75.00 | $5.00 / $25.00 | -67% | Claude Opus 4.5 — Opus tier drops from $15 to $5/M input, 67% reduction |
| Mar 2025 | OpenAI | GPT-5 | LAUNCH | — | $1.25 / $10.00 | — | GPT-5 launch — flagship at $1.25/M, significant drop from GPT-4o |
| Feb 2025 | Gemini 2.0 Flash | LAUNCH | — | $0.10 / $0.40 | — | Gemini 2.0 Flash — $0.10/M input with 1M context, new floor for budget tier | |
| Jan 2025 | DeepSeek | DeepSeek R1 | LAUNCH | — | $0.55 / $2.19 | — | DeepSeek R1 launch — reasoning model at 90% below competitors, market shock |
| Sept 2024 | Gemini 1.5 Pro | PRICE CUT | $7.00 / $21.00 | $3.50 / $10.50 | -50% | Gemini 1.5 Pro price reduction — 50% input cut | |
| Sept 2024 | Mistral | Mistral Small | LAUNCH | — | $0.20 / $0.60 | — | Mistral Small — efficient open-weight model, strong price/performance ratio |
| Sept 2024 | Mistral | Mistral Large | LAUNCH | — | $2.00 / $6.00 | — | Mistral Large — flagship open-weight model at competitive pricing |
| Jul 2024 | OpenAI | GPT-4o mini | LAUNCH | — | $0.15 / $0.60 | — | GPT-4o mini — 60% cheaper than GPT-3.5 Turbo, redefined the budget tier |
| Jun 2024 | Anthropic | Claude 3.5 Sonnet | LAUNCH | — | $3.00 / $15.00 | — | Claude 3.5 Sonnet — significantly better than Claude 3 Sonnet at same price |
| May 2024 | OpenAI | GPT-4o | LAUNCH | — | $2.50 / $10.00 | — | GPT-4o launch — GPT-4 capability at $2.50/M, major price compression |
| Mar 2024 | Anthropic | Claude 3 Opus | LAUNCH | — | $15.00 / $75.00 | — | Claude 3 Opus launch — Anthropic flagship |
| Mar 2024 | Anthropic | Claude 3 Sonnet | LAUNCH | — | $3.00 / $15.00 | — | Claude 3 Sonnet — flagship-adjacent quality at $3/M, set the mid-tier benchmark |
| Mar 2024 | Anthropic | Claude 3 Haiku | LAUNCH | — | $0.25 / $1.25 | — | Claude 3 Haiku — Anthropic enters the budget tier |
| Feb 2024 | GitHub | GitHub Copilot Enterprise | LAUNCH | — | — | — | GitHub Copilot Enterprise — $39/user/month with org-wide features |
| Jan 2024 | OpenAI | GPT-4 Turbo | PRICE CUT | $30.00 / $60.00 | $10.00 / $30.00 | -67% | GPT-4 Turbo — 67% input price reduction |
| Jul 2023 | Anthropic | Claude 2 | LAUNCH | — | $8.00 / $24.00 | — | Claude 2 launch — competitive with GPT-4 at lower price |
| Mar 2023 | OpenAI | GPT-4 | LAUNCH | — | $30.00 / $60.00 | — | GPT-4 launch — set the market benchmark at $30/M input |
| Mar 2023 | OpenAI | GPT-3.5 Turbo | PRICE CUT | $2.00 / $2.00 | $1.50 / $2.00 | -25% | GPT-3.5 Turbo price reduction |
| Feb 2023 | GitHub | GitHub Copilot Business | LAUNCH | — | — | — | GitHub Copilot Business tier — $19/user/month |
| Oct 2022 | GitHub | GitHub Copilot Individual | LAUNCH | — | — | — | GitHub Copilot Individual launches at $10/month per seat — establishes the per-seat coding assistant market |
See how these prices affect your agent costs
Model cost per run, daily, monthly — across every provider.