pplx-embed-v2-context-9b
- Open weights
Prices
0 offerings · prices as listedNo priced offering is listed for this model yet.
Run locally
Weights: perplexity-ai/pplx-embed-v2-context-9b- Parameters 8.4B
- Dense
- Native precision FP32
- Layers 32
- KV cache per token 32 KB
- Config positions 256K (config.json max_position_embeddings)
- Needs
- ≈ 9.1 GB (weights 8.3 + KV cache 0.3 + overhead 0.5)
- Usable
- 21.6 GB (90% of its memory)
- Generation speed
- 55–82 tok/s
- Longest context
- 256K (the model’s limit)
- First token
- about 0.27–0.41 s for a 2K prompt
| Precision | Weights | Needs at 8K | On your hardware | Longest context |
|---|---|---|---|---|
| FP32 (native)maker’s file | 31.3 GB | ≈ 33.1 | ○Needs system memory 3–5 tok/s | — |
| Q8_0 | 8.3 GB | ≈ 9.1 | ●Runs well 55–82 tok/s | 256K |
| FP8 | 7.8 GB | ≈ 8.6 | ●Runs well 58–87 tok/s | 256K |
| Q6_K | 6.4 GB | ≈ 7.2 | ●Runs well 70–105 tok/s | 256K |
| Q5_K_M | 5.6 GB | ≈ 6.3 | ●Runs well 81–121 tok/s | 256K |
| Q5_0 | 5.4 GB | ≈ 6.1 | ●Runs well 83–125 tok/s | 256K |
| Q4_K_M | 4.8 GB | ≈ 5.5 | ●Runs well 94–141 tok/s | 256K |
| Q4_0 | 4.4 GB | ≈ 5.2 | ●Runs well 101–151 tok/s | 256K |
| IQ4_XS | 4.2 GB | ≈ 4.9 | ●Runs well 107–160 tok/s | 256K |
| Q3_K_Mquality drops | 3.8 GB | ≈ 4.6 | ●Runs well 115–173 tok/s | 256K |
| IQ3_XXSquality drops | 3.0 GB | ≈ 3.7 | ●Runs well 145–217 tok/s | 256K |
| Q2_Kquality drops | 2.9 GB | ≈ 3.7 | ●Runs well 147–221 tok/s | 256K |
| IQ2_XXSquality drops | 2.0 GB | ≈ 2.8 | ●Runs well 207–311 tok/s | 256K |
Click a row to choose a precision. “Maker’s file” sizes are the maker’s own files; others are parameters × bits per weight.
Compared with other hardware
Each at the highest precision it runs. Click one for details.
Smallest that runs it (Q4_K_M, 8K): GPU GeForce RTX 3080; unified memory Apple M5 · 16 GB.
Run commands
The maker publishes no files at this precision. Choose its native precision or one of its own builds, or quantise it yourself with llama.cpp.
Memory to fine-tune
Batch 1, 2,048-token sequences, gradient checkpointing; LoRA and QLoRA train rank-16 adapters on the attention projections.
Estimates; facts from the maker’s Hugging Face repositories, as of 2026-11-06. How it’s estimated
Cited scores
From third-party leaderboards, not measured by usNo cited leaderboard covers embedding models yet. MTEB’s results are open (CC0); we will cite them per benchmark, never as a combined score.
Changes
All changes →- New embedding model
Sources
- perplexity-ai/pplx-embed-v2-context-9b config.json (opens in a new tab) huggingface.co · checked Nov 5, 2026
- perplexity-ai/pplx-embed-v2-context-9b model card (opens in a new tab) huggingface.co · checked Nov 5, 2026
- perplexity-ai/pplx-embed-v2-context-9b files (opens in a new tab) huggingface.co · checked Oct 30, 2026
- perplexity-ai/pplx-embed-v2-context-9b model card (opens in a new tab) huggingface.co · checked Oct 30, 2026
Perplexity’s official policies: data use, privacy, terms →
Some values on this page are still waiting to be checked against official pages: a filled dot is verified, a hollow one is not yet (with or without a source on file). How we verify data
Perplexity · United States · perplexity.ai
Related models
Snapshots, the same family and other models from this provider