LANGUAGE MODEL Google/DeepMind

PaLM

Pathways Language Model (540B)

Google's 540-billion-parameter language model, three times larger than GPT-3 and trained on Google's TPU v4 supercomputer. It set state-of-the-art on dozens of benchmarks but was never released as a product, becoming the largest AI model ever to live and die entirely as a research artifact.

Context
2K
Up to 2,048 tokens

Why it matters

PaLM is the cautionary tale every Big Tech AI strategy now studies. Capability without distribution is not a competitive moat. The industry's current obsession with shipping cadence over technical novelty traces directly to this miss.

Core Capabilities

Long Documents
Handles entire codebases, books, and multi-doc RAG.
Generative
Produces images, video, audio, or other media.

Context Window

2k tokens
short prompt
4k Chat 聊天
32k Long docs 长文档
128k Books 整本书
400k Multi-doc 多文档
1M Codebase 整个代码库
10M
2k

Availability

API
Not available
Product / App
Not available
Open Source
Not released
Enterprise

Pricing Model

Research artifact
Not commercially released.
Research

Capability / Performance

Where this model sits relative to the middle 60% of models in the tree. All scores are 0–10 (higher is better).

Lower 20% Upper 80% This model
Reasoning
AA Intelligence Index · scaled to 10
1.7
5.6
1.2
Coding
AA Coding Index · scaled to 10
1.8
4.3
0.7
Context / memory
Context window size · log-scaled
6.0
9.0
0.0
Lower 20% 20th percentile — 20% of models score below this This model Where the current model lands Upper 80% 80th percentile — only 20% of models score above this Percentile boundaries are computed across every model in the tree that reports the underlying benchmark for each capability.

What it feels like

  • Language model from Google — see the linked sources below for benchmark and review coverage

Best use cases

  • General-purpose tasks within Google's deployment footprint
  • See the model spec and sources block for benchmarked use cases

Tools to try

Not ideal for

  • Tasks far outside the modalities listed in this model's spec
  • Workflows where a more recent successor in the same family scores higher

Model Evolution

View full evolution tree →

Chowdhery, A. · Narang, S. · Devlin, J. · Bosma, M. · et al.