Claude Fable 5.2
- Context
- 1M
- Speed
- 102 tok/s
- Size
- Undisclosed
Independent scores, prices, speed and real-world usage for 499 models from 154 labs. Every number links to its source.
Intelligence against price, speed and size for every measured model. Hover any dot to see where it came from.
Artificial Analysis Intelligence Index against blended API price or output speed. Only models with both measurements are plotted.
The four highest Intelligence Index scores, side by side.
Six picks computed from measured results. “Near the top” means within 15% of the best Intelligence Index.
Releases from the last 30 days and the biggest usage movers on OpenRouter this week.
Right column: Intelligence Index, where measured.
Week-over-week change in tokens among the top 15 models on OpenRouter.
The top five on each board. Every board has its own page with the full list.
The model people prefer most in blind side-by-side votes on each arena.
How many of the 92 tracked benchmarks still separate frontier models. Saturated suites no longer tell models apart.
CritPt v2 (Hard): research-level physics.
Leader: Claude Fable 5.2 (76.5%)Terminal-Bench 4.0: autonomous terminal work.
Leader: Kimi K3.1 Pro (67.2%)AIME 2025 competition problems.
Leader: Nova 2.0 Lite (94.3%)Output speed against price and cost to run.
Leader: Celeris-1 (818 tok/s)Each lab ranked by its best measured model, and the coding agents measured end to end.
| Agent | Model used | Index | Cost / task | Time / task |
|---|---|---|---|---|
| Claude Code Anthropic | Claude Fable 5.2 (Xhigh, Default Fallback) | 58.5 | $15.47 | 53 min |
| Google Antigravity Google | Gemini 4 Carbon (High) | 55.5 | $7.13 | 34 min |
| Muse Code Meta | Muse Wonder Full (Max) | 54.5 | — | 34 min |
| Codex OpenAI | GPT-6 Astra (Max) | 53.1 | $7.63 | 26 min |
| Kimi Code CLI Kimi | Kimi K3.1 Pro | 48.6 | $2.94 | 42 min |
Artificial Analysis Coding Agent Index: FrontierCode-AA 2.0, Terminal-Bench 4.1 and SWE-Atlas-QnA.
Scores and release dates come from Artificial Analysis, prices and usage from OpenRouter, benchmark facts from the original papers. Nothing is estimated.