LLM API pricing index, October 2026
List prices, cached-input rates, cache-write surcharges and blended cost per workload for every model the gateway prices, with provider and tier medians.
Across 62 production models from 5 providers in pricing table 2026-10-26, the median blended rate on a 2,000-in / 300-out call is $7.61 per million tokens for frontier models, $3.04 for mid and $0.52 for fast; the cheapest fast model is gpt-5-nano and the cheapest frontier is gemini-2.5-pro.
· Source: TokenZero pricing table, benchmark result files and public audit ledger; CC BY 4.0 with attribution
At a glance
- Models
- 62
- Providers
- OpenAI, Anthropic, Gemini, xAI, DeepSeek
- openai median blended $/1M
- $2.91 (computed)
- anthropic median blended $/1M
- $6.85 (computed)
- gemini median blended $/1M
- $0.98 (computed)
- xai median blended $/1M
- $1.97 (computed)
- deepseek median blended $/1M
- $1.04 (computed)
- Frontier / mid / fast median $/1M
- $7.61 / $3.04 / $0.52 (computed)
- Cheapest per tier
- gemini-2.5-pro / grok-4.3 / gpt-5-nano (computed)
LLM API pricing index, October 2026 (first 62 of 62 rows; full dataset in the downloads)
| model | provider | tier | input_usd_per_1m | output_usd_per_1m | cached_input_usd_per_1m | cache_write_usd_per_1m | cache_discount_pct | blended_usd_per_1m_2000in_300out | cascade_sibling | retires |
|---|---|---|---|---|---|---|---|---|---|---|
| gpt-6-astra | openai | frontier | 10 | 50 | 1 | 12.5 | 90 | 15.22 | gpt-6.1-sol | — |
| gpt-6.1-sol | openai | mid | 2 | 10 | 0.1 | 2.5 | 95 | 3.04 | gpt-6-luna | — |
| gpt-6-sol | openai | mid | 2 | 10 | 0.2 | 2.5 | 90 | 3.04 | gpt-6-luna | — |
| gpt-6-luna | openai | fast | 0.1 | 0.5 | 0.01 | 0.125 | 90 | 0.15 | — | — |
| gpt-5.6-sol | openai | frontier | 4 | 20 | 0.4 | 5 | 90 | 6.09 | gpt-5.6-terra | — |
| gpt-5.6-terra | openai | mid | 2 | 12 | 0.2 | 2.5 | 90 | 3.3 | gpt-5.6-luna | — |
| gpt-5.6-luna | openai | fast | 0.2 | 1.2 | 0.02 | 0.25 | 90 | 0.33 | — | — |
| gpt-5.5 | openai | frontier | 5 | 30 | 0.5 | — | 90 | 8.26 | gpt-5.4-mini | — |
| gpt-5.5-pro | openai | frontier | 30 | 180 | — | — | — | 49.57 | — | — |
| gpt-5.4 | openai | frontier | 2.5 | 15 | 0.25 | — | 90 | 4.13 | gpt-5.4-mini | — |
| gpt-5.4-pro | openai | frontier | 30 | 180 | — | — | — | 49.57 | — | — |
| gpt-5.4-mini | openai | fast | 0.75 | 4.5 | 0.075 | — | 90 | 1.24 | — | — |
| gpt-5.4-nano | openai | fast | 0.2 | 1.25 | 0.02 | — | 90 | 0.34 | — | 2027-04-01 |
| gpt-5.2 | openai | frontier | 1.75 | 14 | 0.175 | — | 90 | 3.35 | gpt-5-mini | — |
| gpt-5.2-pro | openai | frontier | 21 | 168 | — | — | — | 40.17 | — | — |
| gpt-5.1 | openai | mid | 1.25 | 10 | 0.125 | — | 90 | 2.39 | gpt-5-mini | 2027-04-01 |
| gpt-5 | openai | mid | 1.25 | 10 | 0.125 | — | 90 | 2.39 | gpt-5-mini | 2026-12-11 |
| gpt-5-mini | openai | fast | 0.25 | 2 | 0.025 | — | 90 | 0.48 | — | 2026-12-11 |
| gpt-5-nano | openai | fast | 0.05 | 0.4 | 0.005 | — | 90 | 0.1 | — | 2026-12-11 |
| gpt-5-pro | openai | frontier | 15 | 120 | — | — | — | 28.7 | — | 2026-12-11 |
| o3 | openai | mid | 2 | 8 | 0.5 | — | 75 | 2.78 | o4-mini | 2026-12-11 |
| o3-pro | openai | frontier | 20 | 80 | — | — | — | 27.83 | — | 2026-12-11 |
| o3-mini | openai | fast | 1.1 | 4.4 | 0.55 | — | 50 | 1.53 | — | 2026-10-23 |
| o4-mini | openai | fast | 1.1 | 4.4 | 0.275 | — | 75 | 1.53 | — | 2026-10-23 |
| o1 | openai | frontier | 15 | 60 | 7.5 | — | 50 | 20.87 | — | 2026-10-23 |
| gpt-4.1 | openai | mid | 2 | 8 | 0.5 | — | 75 | 2.78 | gpt-4.1-mini | 2027-01-14 |
| gpt-4.1-mini | openai | fast | 0.4 | 1.6 | 0.1 | — | 75 | 0.56 | — | 2027-01-14 |
| gpt-4.1-nano | openai | fast | 0.1 | 0.4 | 0.025 | — | 75 | 0.14 | — | 2026-10-23 |
| gpt-4o | openai | mid | 2.5 | 10 | 1.25 | — | 50 | 3.48 | gpt-4o-mini | 2027-01-14 |
| gpt-4o-mini | openai | fast | 0.15 | 0.6 | 0.075 | — | 50 | 0.21 | — | 2027-01-14 |
| claude-fable-5-1 | anthropic | frontier | 10 | 50 | 0.25 | 12.5 | 98 | 15.22 | claude-opus-5-5 | — |
| claude-fable-5 | anthropic | frontier | 10 | 50 | 1 | 12.5 | 90 | 15.22 | claude-opus-5-5 | — |
| claude-opus-5-5 | anthropic | frontier | 4 | 20 | 0.2 | 5 | 95 | 6.09 | claude-sonnet-5-5 | — |
| claude-opus-5 | anthropic | frontier | 5 | 25 | 0.5 | 6.25 | 90 | 7.61 | claude-sonnet-5-5 | — |
| claude-opus-4-8 | anthropic | frontier | 5 | 25 | 0.5 | 6.25 | 90 | 7.61 | claude-sonnet-5-5 | — |
| claude-opus-4-7 | anthropic | frontier | 5 | 25 | 0.5 | 6.25 | 90 | 7.61 | claude-sonnet-5-5 | — |
| claude-opus-4-6 | anthropic | frontier | 5 | 25 | 0.5 | 6.25 | 90 | 7.61 | claude-sonnet-4-6 | — |
| claude-opus-4-5 | anthropic | frontier | 5 | 25 | 0.5 | 6.25 | 90 | 7.61 | — | — |
| claude-sonnet-5-5 | anthropic | mid | 2 | 10 | 0.1 | 2.5 | 95 | 3.04 | claude-haiku-5-5 | — |
| claude-sonnet-5 | anthropic | mid | 2 | 10 | 0.2 | 2.5 | 90 | 3.04 | claude-haiku-5-5 | — |
| claude-sonnet-4-6 | anthropic | mid | 3 | 15 | 0.3 | 3.75 | 90 | 4.57 | claude-haiku-4-5 | — |
| claude-sonnet-4-5 | anthropic | mid | 3 | 15 | 0.3 | 3.75 | 90 | 4.57 | claude-haiku-4-5 | 2026-11-30 |
| claude-haiku-5-5 | anthropic | fast | 0.1 | 0.5 | 0.01 | 0.125 | 90 | 0.15 | — | — |
| claude-haiku-4-5 | anthropic | fast | 1 | 5 | 0.1 | 1.25 | 90 | 1.52 | — | — |
| gemini-4-argon | gemini | frontier | 2 | 10 | 0.1 | — | 95 | 3.04 | gemini-3.8-flash | — |
| gemini-3.8-flash | gemini | fast | 0.75 | 3.75 | 0.075 | — | 90 | 1.14 | gemini-3.5-flash-lite | — |
| gemini-3.6-flash | gemini | fast | 0.75 | 3.75 | 0.075 | — | 90 | 1.14 | — | — |
| gemini-3.5-flash-lite | gemini | fast | 0.3 | 2.5 | 0.03 | — | 90 | 0.59 | — | — |
| gemini-3.1-pro | gemini | frontier | 2 | 12 | 0.2 | — | 90 | 3.3 | gemini-3.8-flash | — |
| gemini-3.1-flash-lite | gemini | fast | 0.25 | 1.5 | 0.025 | — | 90 | 0.41 | — | — |
| gemini-3-flash | gemini | fast | 0.5 | 3 | 0.05 | — | 90 | 0.83 | — | — |
| gemini-2.5-pro | gemini | frontier | 1.25 | 10 | 0.125 | — | 90 | 2.39 | gemini-2.5-flash | — |
| gemini-2.5-flash | gemini | fast | 0.3 | 2.5 | 0.03 | — | 90 | 0.59 | gemini-2.5-flash-lite | — |
| gemini-2.5-flash-lite | gemini | fast | 0.1 | 0.4 | 0.01 | — | 90 | 0.14 | — | — |
| grok-4.7 | xai | frontier | 2 | 6 | 0.5 | — | 75 | 2.52 | grok-4.3 | — |
| grok-4.6 | xai | frontier | 2 | 6 | 0.5 | — | 75 | 2.52 | grok-4.3 | — |
| grok-4.5 | xai | frontier | 2 | 6 | 0.3 | — | 85 | 2.52 | grok-4.3 | — |
| grok-4.3 | xai | mid | 1.25 | 2.5 | 0.2 | — | 84 | 1.41 | — | — |
| grok-4.20 | xai | mid | 1.25 | 2.5 | 0.2 | — | 84 | 1.41 | — | — |
| grok-build-0.1 | xai | fast | 1 | 2 | 0.2 | — | 80 | 1.13 | — | — |
| deepseek-v4-pro | deepseek | mid | 1.32 | 3.96 | 0.044 | — | 97 | 1.66 | deepseek-flash | — |
| deepseek-flash | deepseek | fast | 0.3 | 1.2 | 0.006 | — | 98 | 0.42 | — | — |
Method and licence
Computed at request time from the same data the gateway runs on; the JSON summary carries every number on this page. Reuse under CC BY 4.0 with a link to https://tokenzero.sh/reports/llm-api-pricing-index-2026-10.
Questions people ask
Which provider has the lowest median price in October 2026?
By median blended rate on the reference call: openai $2.91, anthropic $6.85, gemini $0.98, xai $1.97, deepseek $1.04 per million tokens (table 2026-10-26).
How much more does a frontier model cost than a fast one?
Median blended $7.61 against $0.52 per million tokens, a 14.63x ratio; TokenZero's cascade exists to move simple prompts across that ratio inside one provider with a 99.8% parity floor.
How is the table verified?
Each provider pricing page is fingerprinted and re-checked daily; a drift is flagged on /status and in the admin console before the table is edited. The same table prices every signed receipt.