Model vendor
Leading AI research company
Updated Aug 11, 2026. Data as of 2026-08-11
Default-version specifications and scored-version capability values are labeled separately.
| OP | GPT-5.6 Sol | 95.1 | 80.0% | 1.1M | Closed weights | GPT-5.6 Sol | |
| OP | GPT-5.6 Terra | 87.6 | 80.0% | 1.1M | Closed weights | GPT-5.6 Terra | |
| OP | GPT-5.5 | 84.0 | 80.0% | 1.1M | Closed weights | GPT-5.5 | |
| OP | GPT-5.4 | 77.1 | 80.0% | 1.1M | Closed weights | GPT-5.4 | |
| OP | GPT-5.6 Luna | 76.4 | 80.0% | 1.1M | Closed weights | GPT-5.6 Luna | |
| OP | GPT-5.5 Pro | 75.8 | 40.0% | 1.1M | Closed weights | GPT-5.5 Pro | |
| OP | GPT-5.2 Pro | 71.6 | 60.0% | 400K | Closed weights | GPT-5.2 Pro | |
| OP | GPT-5.2 | 71.4 | 80.0% | 400K | Closed weights | GPT-5.2 | |
| OP | GPT-5.3 Codex | 62.9 | 20.0% | 400K | Closed weights | GPT-5.3 Codex | |
| OP | GPT-5.1 | 62.1 | 60.0% | 400K | Closed weights | GPT-5.1 | |
| OP | GPT-5.1 Instant | 61.3 | 60.0% | 400K | Closed weights | GPT-5.1 Instant | |
| OP | GPT-5.4 mini | 60.5 | 60.0% | 400K | Closed weights | GPT-5.4 mini | |
| OP | GPT-5.2 Codex | 58.9 | 20.0% | 400K | Closed weights | GPT-5.2 Codex | |
| OP | GPT-5.5 Instant | 53.7 | 60.0% | 400K | Closed weights | GPT-5.5 Instant | |
| OP | o3 | 52.6 | 100.0% | 200K | Closed weights | o3 | |
| OP | GPT-5 Medium | 49.8 | 40.0% | 400K | Closed weights | GPT-5 Medium | |
| OP | GPT-5.1 Codex | 49.5 | 20.0% | 400K | Closed weights | GPT-5.1 Codex | |
| OP | GPT-5.4 nano | 49.1 | 60.0% | 400K | Closed weights | GPT-5.4 nano | |
| OP | GPT OSS 20B High | 46.5 | 40.0% | 131.1K | Closed weights | GPT OSS 20B High | |
| OP | o4-mini | 45.0 | 100.0% | 200K | Closed weights | o4-mini | |
| OP | GPT-5 mini | 44.9 | 60.0% | 400K | Closed weights | GPT-5 mini | |
| OP | GPT OSS 120B High | 44.0 | 100.0% | 131.1K | Closed weights | GPT OSS 120B High | |
| OP | GPT-5 Codex | 43.6 | 20.0% | 400K | Closed weights | GPT-5 Codex | |
| OP | GPT-5.3 Chat | 42.2 | 20.0% | 128K | Closed weights | GPT-5.3 Chat | |
| OP | o1 | 36.3 | 80.0% | 200K | Closed weights | o1 | |
| OP | o3-mini | 36.3 | 100.0% | 200K | Closed weights | o3-mini | |
| OP | GPT-4.5 | 36.2 | 100.0% | 128K | Closed weights | GPT-4.5 | |
| OP | o1-pro | 34.2 | 40.0% | 200K | Closed weights | o1-pro | |
| OP | GPT-4.1 | 31.9 | 100.0% | 1M | Closed weights | GPT-4.1 | |
| OP | GPT-5 nano | 31.1 | 60.0% | 400K | Closed weights | GPT-5 nano | |
| OP | GPT-4.1 mini | 24.8 | 100.0% | 1M | Closed weights | GPT-4.1 mini | |
| OP | GPT-4o | 23.9 | 100.0% | 128K | Closed weights | GPT-4o | |
| OP | o1-mini | 16.8 | 80.0% | 128K | Closed weights | o1-mini | |
| OP | GPT-4 Turbo | 14.6 | 80.0% | 128K | Closed weights | GPT-4 Turbo | |
| OP | GPT-4o mini | 8.1 | 80.0% | 128K | Closed weights | GPT-4o mini | |
| OP | GPT-4 | 4.6 | 80.0% | 8.2K | Closed weights | GPT-4 | |
| OP | GPT-4.1 nano | 4.3 | 100.0% | 1M | Closed weights | GPT-4.1 nano | |
| OP | GPT-3.5 Turbo | 0.0 | 80.0% | 16.4K | Closed weights | GPT-3.5 Turbo | |
| OP | GPT-5.1 Codex Mini | 0.0 | 20.0% | 400K | Closed weights | GPT-5.1 Codex Mini | |
| OP | GPT Image 1 | N/A | N/A | N/A | Closed weights | N/A | Not scored |
| OP | GPT Image 1 Mini | N/A | N/A | N/A | Closed weights | N/A | Not scored |
| OP | GPT Image 1.5 | N/A | N/A | N/A | Closed weights | N/A | Not scored |
| OP | GPT Image 2 | N/A | N/A | N/A | Closed weights | N/A | Not scored |
| OP | Sora 2 | N/A | N/A | N/A | Closed weights | N/A | Not scored |
| OP | Sora 2 Pro | N/A | N/A | N/A | Closed weights | N/A | Not scored |
| OP | TTS-1 | N/A | N/A | N/A | Closed weights | N/A | Not scored |
| OP | TTS-1 HD | N/A | N/A | N/A | Closed weights | N/A | Not scored |
| OP | Whisper V1 | N/A | N/A | N/A | Closed weights | N/A | Not scored |
| OP | Whisper Large V3 | N/A | N/A | N/A | Closed weights | N/A | Not scored |
| OP | Whisper Large V3 Turbo | N/A | N/A | N/A | Closed weights | N/A | Not scored |
| OP | o3-pro | N/A | N/A | 200K | Closed weights | Not scored |
Each unique model uses its latest scored version.
Leading OpenAI models by score, price and context window.
Background and model catalog information for OpenAI.
Leading AI research company
Model ownership is separate from API hosting. One model can be offered by multiple API providers with different prices.
Catalog data as of 2026-08-11.
Common questions about OpenAI.
GPT-5.6-Sol currently has the strongest LLMBoard score from this vendor.
The catalog contains 51 independently selectable OpenAI models and 84 versions.
Not necessarily. Vendor identifies the model developer; API providers are tracked separately.
No current default version is marked as open weight.