2 sources, 1 independent of the vendor: centre 8; fewer than 3 sources or none independent, so shown as a band.
- Artificial Analysis · third party
- 10.3 SWE-bench Verified: 72.2 · vendor
- bfcl v4: 70.64 · vendor · not used: no reference models for this benchmark
- claw eval: 51 · vendor · not used: only 0 reference models
- ifbench: 57 · vendor · not used: only 0 reference models
- mrcr 16k 256k: 80.37 · vendor · not used: no reference models for this benchmark
- pinchbench: 87.4 · vendor · not used: most reference models share one score
- tau2 bench: 78.36 · vendor · not used: only 0 reference models
how FTI is computed