What GPU do I need to run allenai/olmOCR-2-7B-1025-FP8?
8.3B parameters, published in F8_E4M3. View on Hugging Face
Set up olmOCR-2-7B-1025-FP8olmOCR-2-7B-1025-FP8 is published by allenai on Hugging Face, with 260,025 downloads and 254 likes to date. It's a Qwen2_5_VLForConditionalGeneration model built for image-text-to-text, published natively in F8_E4M3.
VRAM required & cheapest live GPU fit
Required VRAM = weight size at each precision times a flat 1.2 overhead for activations and fragmentation. The KV-cache grows with context and is not in that factor; it is listed per model below where the architecture is published. Full formula and assumptions: methodology.
| Precision | Weight size | Required VRAM | Cheapest live fit | GPUs needed | Est. $/hr (full fit) |
|---|---|---|---|---|---|
| FP8 (native) | 7.7 GB | 9.3 GB | RTX 4000 SFF Ada | 1 | $0.198/hr |
| INT4 (quantized) | 3.9 GB | 4.6 GB | RTX 3070 | 1 | $0.143/hr |
A GPU is only matched to a row if its hardware supports that precision, and the primary recommendation is always a single-GPU fit when one exists.
INT4 requires a quantized checkpoint actually published for this model, check its Hugging Face page before relying on this row.
Cheapest way to run olmOCR-2-7B-1025-FP8 at its published (F8_E4M3) precision: 1× RTX 4000 SFF Ada, at $0.198/hr per GPU ($0.198/hr total). Quantizing to FP8 or INT4 (rows above) can cost less, but requires a compatible quantized checkpoint to exist for this model.
olmOCR-2-7B-1025-FP8: KV cache by context length
The KV cache is the memory the attention layers hold for every token of context, on top of the weights. It is not part of the flat 1.2x overhead in the table above, grows with context length and with every concurrent request, and is why a long-context deployment needs more VRAM than the table shows.
Grouped-query attention: every layer caches keys and values for a few shared KV heads. Cached per token: 57,344 bytes at 16-bit.
| Context | KV cache, one sequence |
|---|---|
| 4K | 0.22 GB |
| 32K | 1.8 GB |
| 125K (model maximum) | 6.8 GB |
Computed from the layer, head and window counts in the model's own configuration (allenai/olmOCR-2-7B-1025-FP8), at 16 bits per cached value. Each concurrent sequence needs its own cache, on top of the weights.
olmOCR-2-7B-1025-FP8: common questions
Does olmOCR-2-7B-1025-FP8 fit on a 12 GB GPU?
Yes. At FP8 (native) it needs 9.3 GB of VRAM, so a 12 GB card holds it with 2.7 GB to spare. An 8 GB card is not enough for it at FP8 (native).
Is olmOCR-2-7B-1025-FP8 already quantized?
Yes. It is published in FP8, one byte per parameter, so the 9.3 GB figure is already a quantized footprint rather than a full-precision one. Only INT4 goes below it, at 4.6 GB. A GPU without FP8 tensor cores cannot run it as published, which is why cards here are matched on precision support and not on VRAM alone.
How many copies of olmOCR-2-7B-1025-FP8 fit on one RTX 4000 SFF Ada?
2, by VRAM alone. That card carries 20.0 GB and one copy needs 9.3 GB at FP8 (native), on a live rate of $0.198/hr for the whole card. Throughput is not modelled here, so 2 copies is not 2 times the requests served.
What is the least VRAM olmOCR-2-7B-1025-FP8 can run in?
4.6 GB, at INT4 (quantized), which fits a 6 GB card, against 9.3 GB at FP8 (native). That row assumes an INT4 checkpoint has actually been published for this model, so check its Hugging Face page before planning around it.
Weight-to-VRAM math, the fit rules, and how live prices are normalized: full methodology.
More olmOCR models
- olmOCR-2-7B-1025 (8.3B, BF16)
Lineage
Based on Qwen2.5-VL-7B-Instruct.
Alternatives at this size
Other models for image-text-to-text within about a third of olmOCR-2-7B-1025-FP8's 8.3B parameters, from other model lines.
- Qwen3.5-9B (9.7B, BF16)
- llava-1.5-7b-hf (7.1B, F16)
- gemma-3-12b-it (12.2B, BF16)
- UI-TARS-1.5-7B (8.3B, F32)
- Cosmos-Reason2-8B (8.8B, BF16)
More on olmOCR-2-7B-1025-FP8
Fits on a 12 GB GPU at FP8: every model that fits in 12 GB.
Fits on an 8 GB GPU at INT4: every model that fits in 8 GB.
Related reading: H100 pricing and specs, How much VRAM you need for LLMs, Serving LLMs with vLLM, vLLM vs TensorRT-LLM vs SGLang, and Best GPU for LLM inference.
