tinyanalysis

Leaderboard / Ai2 (OLMo)

OLMo 3 7B Think

Ai2 (OLMo) · OLMo 3 · released 2025-11-20 · data checked 2026-11-01

OLMo 3 7B Think is a 7.3B-parameter dense model from Ai2 (OLMo) (United States), released 2025-11-20 under the Apache-2.0 license. It scores 43.4 on the Small LM Index, #17 of 26 small models. It needs about 7.8 GB of memory at 4-bit with an 8K-token context.

Small LM Index
43.4
#17 of 26 scored small models
Likely 38.5–48.3 · from 2 of 6 core benchmarks
Parameters
7.3B
Dense
Price per 1M tokens
—
No hosted API on OpenRouter; self-host
Context window
64K
tokens

Benchmarks

Each bar shows this model's score. The tick marks the best small model on that benchmark. Sources: 5 reported by Ai2 (OLMo). How we label sources.

GPQA Diamond Lab-reported Knowledge46.2
Class best: Qwen3.5 9B 81.7
AIME Lab-reported Math · AIME 202564.6
Class best: Apriel 1.6 15B Thinker 88.0
IFBench Lab-reported Instruction following41.6
Class best: Granite 4.2 8B 79.3
MMLU Lab-reported Knowledge77.8
IFEval Lab-reported Instruction following88.2
Class best: LFM2.5 8B-A1B 91.8

Not reported: estimated

What the index model expects from this model's other scores. Not a measurement.

MMLU-Pro≈ 68
Humanity's Last Exam≈ 5
LiveCodeBench≈ 49
SWE-bench Verified≈ 23

Specifications

Lab
Ai2 (OLMo) (US · United States)
Architecture
Dense transformer
Total parameters
7.3B
Active parameters
All (dense)
Size class
Small
Context window
64K tokens
Input modalities
text
Reasoning
Reasoning (thinking)
License
Apache-2.0
Openness
Fully open: weights, training data and code
Memory to run
~7.8 GB at 4-bit, 8K context (more sizes)
Hugging Face
94.8K downloads (30 days) · 110 likes
Weights on Hugging Face ↗

Compare with

Ranked in

Sources

Something wrong here? Report it.

Run it locally

Uses unsloth/Olmo-3-7B-Think-GGUF (Q4_K_M). Swap the tag for another size: BF16, IQ4_NL, IQ4_XS, Q2_K, Q2_K_L, Q3_K_M, Q3_K_S, Q4_0….

Ollama
ollama run hf.co/unsloth/Olmo-3-7B-Think-GGUF:Q4_K_M
llama.cpp
llama-server -hf unsloth/Olmo-3-7B-Think-GGUF:Q4_K_M
Then open http://localhost:8080 for a chat window.
MLX (Mac)
pip install -U mlx-lm
mlx_lm.chat --model mlx-community/Olmo-3-7B-Think-4bit
vLLM (GPU server)
vllm serve allenai/Olmo-3-7B-Think
Full-precision weights, for a server GPU.

Open in LM Studio

Memory needed: weights, context cache and runtime buffers.

Quantization4K context32K context64K context
16-bit F1617 GB21 GB25 GB
8-bit Q8_011 GB14 GB18 GB
4-bit Q4_K_M7.2 GB11 GB15 GB

Check it against your GPU, Mac or phone →