codellama-70b-instruct
1,098±30
1,068CI1,127
aiqbench · LLM benchmarks
Benchmarks
- Sources
- 8
- Entries tracked
- 22,308
- Last sync
- 19 min19 minutes ago
All sources syncing normally
- Arena · Text· 12 hours ago
- Arena · WebDev· 12 hours ago
- Arena · Vision· 12 hours ago
- Arena · Agent· 12 hours ago
- GPQA Diamond· 19 minutes ago
- SWE-bench· 19 minutes ago
- Terminal-Bench· 19 minutes ago
- ARC-AGI-2· 19 minutes ago
Filter by org:
Released:
One grid, every column. Pick axes in the Axes panel — one axis ranks like a single board, several blend into a weighted composite, all of them read as the index. Rank is assigned before any filter and never renumbers.
| Model | Org | ± CI | Swing | Add to compare | |||||
|---|---|---|---|---|---|---|---|---|---|
| 373 | codellama-70b-instruct | meta | Jan 2024 | — | 1,098 | 1,068–1,127 | #388–825 |
Showing Instruction following exactly as the source published it — its own score and confidence interval, no blending. Select a second axis to build a weighted composite.