OpenGPTAI decision intelligenceAuditable source history
Leaderboard changes & version history
See what moved, what first appeared, and how exact model versions differ—without mixing sources or inventing missing data.Change data is source-specific
Rank movement is calculated only between comparable LMArena Text Style Control overall snapshots. Artificial Analysis and LiveBench remain current-only until OpenGPT has verified historical snapshots for those sources.
This month
Fastest riserFirst appeared in this ranking
Fastest riser
- gpt 5.5openaiUp 1 place · 17 → 16
First appeared in this ranking
- muse spark 1.1meta5
- kimi k3moonshot10
- gpt 5.6 sol xhighopenai11
- gemini 3.6 flashgoogle12
- gemini 3.5 flash highgoogle17
- gemini 3.5 flash mediumgoogle21
- grok 4.5xai33
- claude sonnet 5 highanthropic38
- gemini 3.5 flash litegoogle43
- inklingthinky62
- glm 5v turbozai74
First appearance is not an official product release date.
Compared with Jun 25, 2026
Rank movement
Exact publisher record IDs are compared. New entries are not treated as rising from zero and can push existing records down, so a lower rank does not prove lower model quality.379
| Model record | Provider | Previous | Current | Change | Rating change |
|---|---|---|---|---|---|
| claude fable 5 | anthropic | 1 | 1 | No rank change | -0.86 |
| claude opus 4 6 thinking | anthropic | 2 | 2 | No rank change | +1.75 |
| claude opus 4 7 thinking | anthropic | 3 | 3 | No rank change | -0.26 |
| claude opus 4 6 | anthropic | 4 | 4 | No rank change | -1.31 |
| muse spark 1.1 | meta | — | 5 | First appeared in this snapshot | — |
| claude opus 4 7 | anthropic | 5 | 6 | Down 1 place | -0.14 |
| muse spark | meta | 6 | 7 | Down 1 place | +0.27 |
| gemini 3.1 pro preview | 7 | 8 | Down 1 place | -0.27 | |
| gemini 3 pro | 8 | 9 | Down 1 place | -0.06 | |
| kimi k3 | moonshot | — | 10 | First appeared in this snapshot | — |
| gpt 5.6 sol xhigh | openai | — | 11 | First appeared in this snapshot | — |
| gemini 3.6 flash | — | 12 | First appeared in this snapshot | — | |
| claude opus 4 8 thinking | anthropic | 9 | 13 | Down 4 places | -0.13 |
| gpt 5.5 high | openai | 10 | 14 | Down 4 places | +0.47 |
| gpt 5.4 high | openai | 12 | 15 | Down 3 places | +0.03 |
| gpt 5.5 | openai | 17 | 16 | Up 1 place | +1.50 |
| gemini 3.5 flash high | — | 17 | First appeared in this snapshot | — | |
| gpt 5.2 chat latest 20260210 | openai | 14 | 18 | Down 4 places | +0.18 |
| qwen3.7 max preview | alibaba | 16 | 19 | Down 3 places | +0.31 |
| grok 4.20 beta1 | xai | 18 | 20 | Down 2 places | -0.10 |
| gemini 3.5 flash medium | — | 21 | First appeared in this snapshot | — | |
| claude opus 4 8 | anthropic | 11 | 22 | Down 11 places | -5.62 |
| gpt 5.5 instant | openai | 22 | 23 | Down 1 place | +0.82 |
| grok 4.20 beta 0309 reasoning | xai | 15 | 24 | Down 9 places | -2.33 |
| gemini 3 flash | 20 | 25 | Down 5 places | -0.11 | |
| claude opus 4 5 20251101 thinking 32k | anthropic | 21 | 26 | Down 5 places | +0.02 |
| claude sonnet 4 6 | anthropic | 23 | 27 | Down 4 places | +0.33 |
| grok 4.20 multi agent beta 0309 | xai | 24 | 28 | Down 4 places | -0.14 |
| glm 5.1 | zai | 19 | 29 | Down 10 places | -3.77 |
| glm 5.2 (max) | zai | 25 | 30 | Down 5 places | -0.67 |
Why only one ranking has movement data
The current June and July LMArena snapshots use the same documented method epoch. Earlier archive periods are labeled separately. OpenGPT has not frozen a comparable prior edition for the other ranking sources, so no movement is shown for them.
Version differencesCompare two catalog model IDs from one model family. Provider IDs, sources and published ranking signals are shown; OpenGPT catalog classification is labeled separately.
| Version differences | Claude Fable 5 | Claude Opus 4.7 |
|---|---|---|
| Official model ID | claude-fable-5 | claude-opus-4-7 |
| Directory status | General availability | General availability |
| OpenGPT catalog role | Flagship | Previous |
| Current public ranking signals | User preference 1 · 1507 · Intelligence index 3 · 62 · Objective tasks 1 · 83.4 | User preference 6 · 1494 |
| Provider source | Anthropic ↗ | Anthropic ↗ |
| Model details | Model details → | Model details → |
No verified feature-difference data is available. OpenGPT does not infer speed, price, context length, or quality.
Historical snapshotsOne frozen source snapshot is preserved for each calendar month. The manifest publishes a SHA-256 file check, and the generator refuses silent replacement.
| Month | Source publication | Method epoch | Entries | File check | Download JSON |
|---|---|---|---|---|---|
| July 2026 Current | Method epoch 3 | 378 | 30a2b9915774… | Download JSON ↓ | |
| June 2026 | Method epoch 3 | 368 | a06ed961a983… | Download JSON ↓ | |
| May 2026 | Method epoch 3 | 360 | 057c9e12679b… | Download JSON ↓ | |
| April 2026 | Method epoch 3 | 354 | dca5faa06859… | Download JSON ↓ | |
| March 2026 | Method epoch 3 | 336 | c258a21cfd6f… | Download JSON ↓ | |
| February 2026 | Method epoch 3 | 319 | 4f21abee9ac0… | Download JSON ↓ | |
| January 2026 | Method epoch 3 | 302 | b252f774df59… | Download JSON ↓ | |
| December 2025 | Method epoch 3 | 291 | 75a5e543c2b0… | Download JSON ↓ | |
| November 2025 | Method epoch 3 | 276 | cf161eb704e7… | Download JSON ↓ | |
| October 2025 | Method epoch 3 | 259 | f9589f5985c3… | Download JSON ↓ | |
| September 2025 | Method epoch 3 | 248 | ba97765955d4… | Download JSON ↓ | |
| August 2025 | Method epoch 3 | 238 | 28fbbd1445b0… | Download JSON ↓ |
Earlier records were imported later from LMArena’s official full-history dataset. They are source snapshots, not OpenGPT editions claimed to have been published at those earlier dates.