kimi-k2-0905-preview vs llama-3.1-8b-instruct

Capability for the money

Not the same tier: kimi-k2-0905-preview leads llama-3.1-8b-instruct by 50.8 points on hard reasoning (GPQA Diamond), for 5.4× the price.

1. Following instructions

   extraction, classification, sticking to a format

   IFBench
   kimi-k2-0905-preview41.7%

   llama-3.1-8b-instruct28.6%

   kimi-k2-0905-preview +13.1

2. Coding

   writing and fixing code

   LiveCodeBench
   kimi-k2-0905-preview61.0%

   llama-3.1-8b-instruct11.6%

   kimi-k2-0905-preview +49.4

3. Tool use

   multi-step agent work against real tools

   τ²-bench
   kimi-k2-0905-preview73.4%

   llama-3.1-8b-instruct16.4%

   kimi-k2-0905-preview +57.0

4. Hard reasoning

   graduate-level science and multi-step analysis

   GPQA Diamond
   kimi-k2-0905-preview76.7%

   llama-3.1-8b-instruct25.9%

   kimi-k2-0905-preview +50.8

Averaged over the benchmarks both models sat, llama-3.1-8b-instruct returns 103.1 points of pass rate per dollar of blended price against 58.8 — 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. And a benchmark is not your workload — it narrows the shortlist, it does not pick for you.