claude-fable-5 vs gpt-5.6-sol

Capability for the money

claude-fable-5 is 1.8× dearer for 13.4 points on tool use (τ²-bench) — worth it only if that is your bottleneck.

1. Following instructions

   extraction, classification, sticking to a format

   IFBench
   claude-fable-563.5%

   gpt-5.6-sol72.7%

   gpt-5.6-sol +9.2

2. Tool use

   multi-step agent work against real tools

   τ²-bench
   claude-fable-598.5%

   gpt-5.6-sol85.1%

   claude-fable-5 +13.4

3. Hard reasoning

   graduate-level science and multi-step analysis

   GPQA Diamond
   claude-fable-592.6%

   gpt-5.6-sol94.1%

   gpt-5.6-sol +1.5

Averaged over the benchmarks both models sat, gpt-5.6-sol returns 7.5 points of pass rate per dollar of blended price against 4.2 — the ratio that decides whether the dearer model earns its rate.

Each figure is a pass rate: the share of that test set the model answered correctly, measured by a third party rather than self-reported by the lab. Rows are different tests of different difficulty, so compare a model against the other model, not one row against the next. No published coding score for one of these models, so that lens is omitted. And a benchmark is not your workload — it narrows the shortlist, it does not pick for you.