Compare
Compare by Provider
Pick two to four providers and every model they publish is ranked together — not the providers side by side, but their whole ranges interleaved on one capability table and one quality-versus-cost plot. A figure no source published stays explicitly unavailable.
Provider set
Check two to 4 providers below, then press Compare — every model they publish is ranked together, and the URL and the table only update once you do.
Compare by Provider
59 models across 2 providers, on the same seven capability axes the leaderboard ranks. Choose a column header to sort; expand a row for its benchmark tasks.
OpenAI- Models
- 39
- Scored
- 39
- Best overall
- 82.81GPT-6 Astra
- Blended $ / 1M
- $0.055 – $262.50
Anthropic- Models
- 20
- Scored
- 20
- Best overall
- 84.74Claude Fable 5.1
- Blended $ / 1M
- $0.50 – $30.00
Evidence key: ObservedUnavailable
| Compare | Details | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| 1Claude Fable 5.1Anthropic · Proprietary | 84.74 | 60.8 | 90.4 | 97.5 | 79.4 | Unavailable | Unavailable | Unavailable | ||
| 2GPT-6 AstraOpenAI · Proprietary | 82.81 | 87.1 | 75.2 | 86.9 | 89.5 | 82.9 | Unavailable | 85 | ||
| 3Claude Opus 5Anthropic · Proprietary | 81.87 | 85.6 | 89.7 | 97.5 | 75.7 | 88.8 | Unavailable | Unavailable | ||
| 4Claude Fable 5Anthropic · Proprietary | 81.41 | 89.1 | 89.5 | 87.3 | 77.6 | 62.6 | 77 | Unavailable | ||
| 5GPT-5.6 SolOpenAI · Proprietary | 80.46 | 93.6 | 61.9 | 84.7 | 70.4 | 87.6 | 87.7 | 96.8 | ||
| 6Claude Opus 4.8Anthropic · Proprietary | 72.17 | 76.6 | 73.1 | 88.7 | 55.5 | 87.7 | 74 | 65.2 | ||
| 7GPT-5.5OpenAI · Proprietary | 71.86 | 77.2 | 51.8 | 80.8 | 64 | 71.4 | 91.9 | 69.4 | ||
| 8GPT-5.6 TerraOpenAI · Proprietary | 71.13 | 87.9 | 61.3 | 77.2 | 63.7 | 77.2 | 85.7 | 96.8 | ||
| 9GPT-5.4OpenAI · Proprietary | 70.88 | 72 | 42.8 | 76.7 | 57.2 | 69.3 | 89.2 | 64.4 | ||
| 10Claude Opus 4.7Anthropic · Proprietary | 70.6 | 38.7 | 65.4 | 81.2 | Unavailable | Unavailable | Unavailable | 60.6 | ||
| 11Claude Sonnet 5Anthropic · Proprietary | 69.68 | 82.4 | 59.5 | 82.8 | 77.4 | 77.4 | Unavailable | Unavailable | ||
| 12Claude Opus 4.7 (Adaptive)Anthropic · Proprietary | 69.31 | 64.7 | 62.8 | 82.1 | 49.7 | 50.1 | Unavailable | Unavailable | ||
| 13Claude Opus 4.6Anthropic · Proprietary | 69.21 | 63.5 | 50.3 | 81.6 | 67.1 | 59.6 | 51 | 58.5 | ||
| 14GPT-5.2 ProOpenAI · Proprietary | 68.99 | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | ||
| 15GPT-5.2OpenAI · Proprietary | 64.86 | 23.1 | 46.5 | 77.3 | 53.7 | 66.3 | 91.2 | 57.4 | ||
| 16GPT-5.6 LunaOpenAI · Proprietary | 64.77 | 81.2 | 53.7 | 74.8 | 51.2 | 67.1 | Unavailable | 94.2 | ||
| 17GPT-5.3 CodexOpenAI · Proprietary | 64.03 | 62.8 | 52 | 79.5 | 78.3 | 75.7 | 91.2 | Unavailable | ||
| 18GPT-5.1OpenAI · Proprietary | 63.18 | Unavailable | Unavailable | 63.6 | 76 | 71.3 | 87.9 | 49 | ||
| 19Claude Sonnet 4.6Anthropic · Proprietary | 62.86 | 50.5 | 61.9 | 76.8 | 67.9 | 54.1 | 46.6 | 48.8 | ||
| 20Claude Opus 4.6 (Adaptive)Anthropic · Proprietary | 62.23 | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | ||
| 21GPT-5.5 ProOpenAI · Proprietary | 61.9 | 80.6 | Unavailable | 82.8 | Unavailable | Unavailable | Unavailable | 70.1 | ||
| 22GPT-5.4 miniOpenAI · Proprietary | 61.05 | 55.5 | 56 | 58.3 | 73.9 | 57.3 | 88.5 | 44.3 | ||
| 23GPT-5.4 ProOpenAI · Proprietary | 60.49 | 79.3 | Unavailable | 84.8 | 70.2 | Unavailable | Unavailable | 68.8 | ||
| 24GPT-5.4 nanoOpenAI · Proprietary | 59.42 | 18.5 | 62.6 | 48.2 | 73.7 | 23.8 | 91.9 | 43.8 | ||
| 25Claude Opus 4.5Anthropic · Proprietary | 58.42 | 47.3 | 50.2 | 60.3 | 72.1 | 23.5 | 30.6 | 57.9 | ||
| 26GPT-5.2 InstantOpenAI · Proprietary | 57.05 | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | ||
| 27GPT-5.3 InstantOpenAI · Proprietary | 56.95 | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | ||
| 28GPT-5 (high)OpenAI · Proprietary | 56.7 | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | ||
| 29GPT-5.3-Codex-SparkOpenAI · Proprietary | 54.99 | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | ||
| 30Claude Sonnet 4.5Anthropic · Proprietary | 54.01 | 33.7 | 59.6 | 72.9 | 19.1 | Unavailable | Unavailable | 34.6 | ||
| 31GPT-5.2-CodexOpenAI · Proprietary | 53.95 | Unavailable | 67.2 | 67.2 | 77.6 | 72.5 | 92.4 | Unavailable | ||
| 32Claude Opus 4.5 ThinkingAnthropic · Proprietary | 53.05 | Unavailable | Unavailable | 73 | 74.1 | 69.1 | 68.5 | Unavailable | ||
| 33Claude Haiku 4.5Anthropic · Proprietary | 52.89 | Unavailable | 30.7 | 53.7 | Unavailable | Unavailable | Unavailable | 28.8 | ||
| 34GPT-5.1-Codex-MaxOpenAI · Proprietary | 52.6 | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | ||
| 35GPT-5 (medium)OpenAI · Proprietary | 49.62 | Unavailable | Unavailable | 65.5 | 73.2 | 69.5 | 84.9 | Unavailable | ||
| 36o4-mini (high)OpenAI · Proprietary | 48.2 | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | 43 | ||
| 37GPT-5.1-CodexOpenAI · Proprietary | 48.16 | Unavailable | Unavailable | 65.9 | 68.6 | 66.9 | 84.2 | Unavailable | ||
| 38o3-proOpenAI · Proprietary | 47.84 | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | Unavailable | ||
| 39o3OpenAI · Proprietary | 47.58 | Unavailable | Unavailable | 64.6 | 72.3 | 63.3 | 86 | 37.7 | ||
| 40o3-miniOpenAI · Proprietary | 46.72 | Unavailable | 23 | 66.8 | Unavailable | Unavailable | Unavailable | Unavailable |
2 of these rows are marked self-hosted by the source: open-weight models with no first-party hosted token rate. 2 of them show a hosted rate matched from another provider; the rest report an unavailable price rather than zero.
Quality versus cost
Each point is one of these providers' models with both an overall score and complete token prices. Cost is logarithmic by default; untick for a linear axis.
Lowest blended cost
Up to twelve of these providers' models, ranked on the 75% input / 25% output blend.
2 of these rows are marked self-hosted by the source: open-weight models with no first-party hosted token rate. 2 of them show a hosted rate matched from another provider; the rest report an unavailable price rather than zero.
0 of 4 selected