tokeven38 models across 3 tiers
GPT API pricing in 2026
OpenAI has models across three tiers: flagship (GPT 5.5, GPT 5.6 Sol), mid-tier (GPT 5.4, GPT 5.6 Terra), and fast (GPT 5.6 Luna). The GPT 5.6 family launched with 1M context windows across all tiers. Every model has a 90% cache read discount. The GPT 5.6 family charges a 1.25x cache write surcharge; GPT 5.5 and GPT 5.4 are free to write.
Models
38
Cheapest input
$0.05/MTok
Cache discount
90% off
Max context
1.1M
All GPT models
| Model | Tier | Input $/MTok | Output $/MTok | Context | Best task | Released |
|---|---|---|---|---|---|---|
| GPT 5 Nano | Fast / Budget | $0.05 | $0.40 | 272K | Not benchmarked | — |
| GPT 4.1 Nano | Fast / Budget | $0.10 | $0.40 | 1.0M | Not benchmarked | — |
| GPT 6 Luna | Fast / Budget | $0.10 | $0.50 | 922K | Reasoning (82/100) | — |
| GPT 4o Mini | Fast / Budget | $0.15 | $0.60 | 128K | Not benchmarked | — |
| GPT 5.6 Luna | Fast / Budget | $0.20 | $1.20 | 922K | Classification (91/100) | Jul 2026 |
| GPT 5.4 Nano | Fast / Budget | $0.20 | $1.25 | 272K | Reasoning (81/100) | — |
| GPT 5 Mini | Fast / Budget | $0.25 | $2.00 | 272K | Not benchmarked | — |
| GPT 4.1 Mini | Fast / Budget | $0.40 | $1.60 | 1.0M | Not benchmarked | — |
| GPT 3.5 Turbo | Fast / Budget | $0.50 | $1.50 | 16K | Not benchmarked | — |
| GPT 3.5 Turbo (0125) | Fast / Budget | $0.50 | $1.50 | 16K | Not benchmarked | — |
| GPT 5.4 Mini | Fast / Budget | $0.75 | $4.50 | 272K | Extraction (75/100) | — |
| o3 Mini | Fast / Budget | $1.10 | $4.40 | 200K | Not benchmarked | — |
| o4 Mini | Fast / Budget | $1.10 | $4.40 | 200K | Not benchmarked | — |
| GPT 5 | Mid-tier | $1.25 | $10.00 | 272K | Not benchmarked | — |
| GPT 5.1 | Mid-tier | $1.25 | $10.00 | 272K | Not benchmarked | — |
| GPT 5.2 | Mid-tier | $1.75 | $14.00 | 272K | Not benchmarked | — |
| GPT 5.6 Terra | Mid-tier | $2.00 | $12.00 | 922K | Q&A (92/100) | Jul 2026 |
| GPT 4.1 | Mid-tier | $2.00 | $8.00 | 1.0M | Not benchmarked | — |
| GPT 6 Sol | Flagship | $2.00 | $10.00 | 922K | Reasoning (89/100) | — |
| o3 | Mid-tier | $2.00 | $8.00 | 200K | Not benchmarked | — |
| GPT 6.1 Sol | Flagship | $2.00 | $10.00 | 922K | Not benchmarked | Sep 2026 |
| GPT 5.4 | Mid-tier | $2.50 | $15.00 | 1.1M | Q&A (91/100) | Mar 2026 |
| GPT 4o | Mid-tier | $2.50 | $10.00 | 128K | Not benchmarked | — |
| GPT 5.6 Sol | Flagship | $4.00 | $20.00 | 922K | Code Review (92/100) | Jul 2026 |
| GPT 5.5 | Flagship | $5.00 | $30.00 | 1.1M | Code Review (90/100) | Mar 2026 |
| GPT 4o (2024-05-13) | Flagship | $5.00 | $15.00 | 128K | Not benchmarked | — |
| GPT 4 Turbo (2024-04-09) | Flagship | $10.00 | $30.00 | 128K | Not benchmarked | — |
| GPT 4 Turbo | Flagship | $10.00 | $30.00 | 128K | Not benchmarked | — |
| GPT 6 Astra | Flagship | $10.00 | $50.00 | 922K | Reasoning (93/100) | — |
| GPT 5 Pro | Flagship | $15.00 | $120.00 | 400K | Not benchmarked | — |
| o1 | Flagship | $15.00 | $60.00 | 200K | Not benchmarked | — |
| o3 Pro | Flagship | $20.00 | $80.00 | 200K | Not benchmarked | — |
| GPT 5.2 Pro | Flagship | $21.00 | $168.00 | 272K | Not benchmarked | — |
| GPT 4 (0613) | Flagship | $30.00 | $60.00 | 8K | Not benchmarked | — |
| GPT 4 | Flagship | $30.00 | $60.00 | 8K | Not benchmarked | — |
| GPT 5.4 Pro | Flagship | $30.00 | $180.00 | 1.1M | Not benchmarked | — |
| GPT 5.5 Pro | Flagship | $30.00 | $180.00 | 1.1M | Not benchmarked | — |
| o1 Pro | Flagship | $150.00 | $600.00 | 200K | Not benchmarked | — |
GPT 5 Nano: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
GPT 4.1 Nano: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
GPT 6 Luna: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Benchmarked from the LiveBench 2026-06-25 release (2026-W26) in the first rankings sync.
GPT 4o Mini: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
GPT 5.4 Nano: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Benchmarked from the LiveBench 2026-06-25 release (2026-W26) in the first rankings sync.
GPT 5 Mini: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
GPT 4.1 Mini: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
GPT 3.5 Turbo: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
GPT 3.5 Turbo (0125): Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
GPT 5.4 Mini: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Benchmarked from the LiveBench 2026-06-25 release (2026-W26) in the first rankings sync.
o3 Mini: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
o4 Mini: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
GPT 5: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
GPT 5.1: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
GPT 5.2: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
GPT 4.1: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
GPT 6 Sol: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Benchmarked from the LiveBench 2026-06-25 release (2026-W26) in the first rankings sync.
o3: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
GPT 6.1 Sol: Added by the model-onboarding job on 2026-09-30. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date is the day Tokeven first saw it (2026-09-30) because the provider publishes none in a machine-readable source. Not yet benchmarked, so never recommended.
GPT 4o: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
GPT 4o (2024-05-13): Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
GPT 4 Turbo (2024-04-09): Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
GPT 4 Turbo: Alias of gpt-4-turbo-2024-04-09 (same prices). Added 2026-10-13 for integrations that report the requested model id. Not benchmarked, so never recommended.
GPT 6 Astra: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Benchmarked from the LiveBench 2026-06-25 release (2026-W26) in the first rankings sync.
GPT 5 Pro: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
o1: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
o3 Pro: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
GPT 5.2 Pro: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
GPT 4 (0613): Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
GPT 4: Alias of gpt-4-0613 (same prices). Added 2026-10-13 for integrations that report the requested model id. Not benchmarked, so never recommended.
GPT 5.4 Pro: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
GPT 5.5 Pro: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
o1 Pro: Backfilled on 2026-09-23. Prices confirmed by the provider pricing page and the LiteLLM cross-check. Release date not published in a machine-readable source. Not yet benchmarked, so never recommended.
GPT cost by use case
Estimated cost per API call across common workloads, using default token counts for each use case.
GPT quality benchmarks by task
Confidence scores (0-100) across all task types. Higher is better.
Confidence scores are a third-party benchmark snapshot from week 26 of 2026. This snapshot is 17 weeks old, so treat the scores as directional and validate against your own evals.
| Task | Nano | Nano | Luna | Mini | Luna | Nano | Mini | Mini | Turbo | (0125) | Mini | Mini | Mini | 5 | 5.1 | 5.2 | Terra | 4.1 | Sol | o3 | Sol | 5.4 | 4o | Sol | 5.5 | (2024-05-13) | (2024-04-09) | Turbo | Astra | Pro | o1 | Pro | Pro | (0613) | 4 | Pro | Pro | Pro |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Code Generation | - | - | 62/100 | - | 62/100 | 56/100 | - | - | - | - | 54/100 | - | - | - | - | - | 64/100 | - | 64/100 | - | - | 63/100 | - | 67/100 | 65/100 | - | - | - | 67/100 | - | - | - | - | - | - | - | - | - |
| Code Review | - | - | - | - | 68/100 | - | - | - | - | - | - | - | - | - | - | - | 84/100 | - | - | - | - | 82/100 | - | 92/100 | 90/100 | - | - | - | - | - | - | - | - | - | - | - | - | - |
| Summarization | - | - | 53/100 | - | 57/100 | 67/100 | - | - | - | - | 58/100 | - | - | - | - | - | 64/100 | - | 69/100 | - | - | 67/100 | - | 70/100 | 70/100 | - | - | - | 76/100 | - | - | - | - | - | - | - | - | - |
| Q&A | - | - | - | - | 85/100 | - | - | - | - | - | - | - | - | - | - | - | 92/100 | - | - | - | - | 91/100 | - | 91/100 | 90/100 | - | - | - | - | - | - | - | - | - | - | - | - | - |
| Extraction | - | - | 75/100 | - | 74/100 | 74/100 | - | - | - | - | 75/100 | - | - | - | - | - | 75/100 | - | 77/100 | - | - | 76/100 | - | 75/100 | 78/100 | - | - | - | 79/100 | - | - | - | - | - | - | - | - | - |
| Reasoning | - | - | 82/100 | - | 86/100 | 81/100 | - | - | - | - | 71/100 | - | - | - | - | - | 91/100 | - | 89/100 | - | - | 88/100 | - | 92/100 | 90/100 | - | - | - | 93/100 | - | - | - | - | - | - | - | - | - |
| Classification | - | - | - | - | 91/100 | - | - | - | - | - | - | - | - | - | - | - | 89/100 | - | - | - | - | 88/100 | - | 86/100 | 85/100 | - | - | - | - | - | - | - | - | - | - | - | - | - |
| Creative Writing | - | - | 63/100 | - | 69/100 | 68/100 | - | - | - | - | 67/100 | - | - | - | - | - | 67/100 | - | 67/100 | - | - | 79/100 | - | 76/100 | 73/100 | - | - | - | 74/100 | - | - | - | - | - | - | - | - | - |
Recent GPT pricing changes
GPT 6.1 Sol input launched
$2.00/MTok at launch
GPT 5.6 Sol input decreased
$5.00 to $4.00/MTok
GPT 5.6 Sol output decreased
$30.00 to $20.00/MTok
GPT 5.6 Terra input decreased
$2.50 to $2.00/MTok
GPT 5.6 Terra output decreased
$15.00 to $12.00/MTok
GPT 5.6 Luna input decreased
$1.00 to $0.20/MTok
Compare GPT with other providers
Head-to-head comparisons
Track your GPT spend
See exactly how much you spend on GPT models and where you can save. Two-line wrapper install, no API keys shared.