The Aggregate: unified LLM rankings, daily
Best AI you can use today: Claude Fable 5.2.
Rankings for 10554 models from connected benchmark evidence. Rankings combine measured results using benchmark predictive accuracy and calibration uncertainty, refreshed daily.
Top models by Aggregate Rating
| Rank | Model | Provider | Rating | Benchmarks |
|---|---|---|---|---|
| 1 | Claude Fable 5.2 | Anthropic | 1997 ± 7 | 492 |
| 2 | GPT-6.2 Astra | OpenAI | 1983 ± 29 | 108 |
| 3 | Gemini 4 Carbon | 1969 ± 7 | 478 | |
| 4 | GPT-6 Astra Pro Low coverage | OpenAI | 1965 ± 7 | 23 |
| 6 | GPT-6.2 Sol | OpenAI | 1954 ± 27 | 107 |
| 7 | Claude Opus 5.5 | Anthropic | 1953 ± 4 | 1671 |
| 8 | GPT-6 Astra | OpenAI | 1949 ± 4 | 2161 |
| 9 | GPT-6.1 Pro Sol Low coverage | OpenAI | 1946 ± 10 | 21 |
| 12 | GPT-6.1 Sol | OpenAI | 1934 ± 5 | 1356 |
| 13 | Gemini 4 Hydrogen | 1927 ± 8 | 226 |
Which AI is worth paying for?
Plans ranked by what a month of use is worth on the API, per dollar of price.
| Plan | Value per dollar | Price | Best use of the allowance | Verdict |
|---|---|---|---|---|
| Claude Ultra | 28× | USD 200/mo | Claude Sonnet 5.5 + Claude Fable 5.2, 63B tokens a month, API worth $5,640 | Ten times Pro's weekly Claude usage without a five-hour limit once the account is verified, plus a separate monthly API budget; credits require claiming and cannot extend interactive plan limits. |
| Claude Pro | 27× | USD 20/mo | Claude Sonnet 5.5, 6.2B tokens a month, API worth $539 | Generous Opus allowance for regular chat and coding in one subscription. |
| Claude Max | 27× | USD 100/mo | Claude Sonnet 5.5 + Claude Fable 5.2, 30B tokens a month, API worth $2,670 | Higher Claude usage plus a separate monthly API budget for custom apps and agents; credits require claiming and cannot extend interactive plan limits. |
| ChatGPT Pro 500 | 19× | USD 500/mo | GPT-6.1 Sol, 55B tokens a month, API worth $9,616 | The largest of the three Pro allowances and the only Pro tier with Ultrafast; per dollar it buys about what Pro 200 does, so pay for it when Pro 200 runs out or Ultrafast speed matters. |
| ChatGPT Pro 200 | 19× | USD 200/mo | GPT-6.1 Sol, 22B tokens a month, API worth $3,735 | More agent headroom without a five-hour limit; OpenAI canceled the September reduction, so every Pro 200 subscription has the higher allowance again. |
| ChatGPT Pro 100 | 19× | USD 100/mo | GPT-6.1 Sol, 11B tokens a month, API worth $1,868 | More agent capacity without a five-hour limit, with the broader ChatGPT toolset. |
| OpenCode Go | 16× | USD 10/mo | Muse Spark 1.3, 13B tokens a month, API worth $156 | Low-cost coding access with explicit model-dependent token pools; rolling limits, context prices and contributor terms affect usable capacity. |
| ChatGPT Plus | 8.7× | USD 20/mo | GPT-6.1 Sol, 1.0B tokens a month, API worth $173 | A strong all-round bundle; sustained agent work can outgrow its limits. |
Latest leaderboard changes
- Gemini 4 Hydrogen took the lead on Design Arena (ASCII Art) (1473 vs 1357 by Claude Opus 5.5)
- Qwen 4 Max took the lead on AA APEX-Agents (59.14 vs 47.05 by Gemini 3.5 Flash (High))
- Gemini 4 Hydrogen took the lead on LLM Stats (Agents' Last Exam) (71.6 vs 61.6 by Muse Ember)
- Qwen 4 Max took the lead on LLM Stats (APEX-Agents) (72.1 vs 67.2 by Kimi K3.1 Pro)
- qwen-4 (ALE Run: harness=Codex; effort=xhigh; source-model=qwen-4) took the lead on Agents' Last Exam v1.1 - Codex Pass Rate (46.72 vs 42.62 by GPT-6 Astra (ALE Run: harness=Codex; effort=max; source-model=gpt-6-astra))
- Qwen 4 Max took the lead on LLM Stats (Job Bench) (68.5 vs 64.9 by Muse Spark 1.3)
- LimiX-2-1016 (default) took the lead on TabArena All Tasks (93.2 vs 89.6 by LimiX-2 (default))
- Qwen 4 Max took the lead on EnterpriseOps-Gym-AA - Calendar (53.85 vs 51.28 by Claude Opus 5 (Max))
Explore: all models · benchmark catalog · capability trends · daily changes
Interactive version: theaggregate.ai/ · How It Works · Data refreshed daily, snapshot 2026-11-17.