OpenThinker-32B
Open-data reasoning model from a fully open curation pipeline.
About
OpenThinker-32B is the flagship of the Open Thoughts project, which set out to prove that strong reasoning models can be built with fully open data curation. Released in February 2025 on the OpenThoughts-114k dataset, it was positioned by its team as the best open-data reasoning model of its moment, and it became one of the most downloaded reasoning models on Ollama.
Being an early-2025 reasoning model, it has since been overtaken by newer RL-trained models on hard math and code benchmarks. Its lasting value is the open pipeline: dataset, training recipe, and evaluation harness are all public, which makes it a good base for reasoning research rather than a frontier pick.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
- Positioned as best open-data reasoning model at release on math and code benchmarks (team-reported, Feb 2025)
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 7/10
Scored 7/10 (verified tier, metrics-v1). Evidence: Open Thoughts project, Feb 2025; fully open data curation pipeline, dataset and eval harness public; Strong early adoption (400k+ Ollama downloads by Feb 2025) but overtaken by newer RL-trained reasoning models; Maintenance is community-driven; no successor flagship from the same team in 2025-2026; Benchmark positioning is team-reported from its release moment. License: Apache-2.0 per the Hugging Face page.
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 7.0/10
- Maintenance · recency, releases, issue load
- 6.0/10
- Openness · license + docs
- 9.0/10
- Manual review · sanity read + red-flag check
- 7.0/10
- · Open Thoughts project, Feb 2025; fully open data curation pipeline, dataset and eval harness public
- · Strong early adoption (400k+ Ollama downloads by Feb 2025) but overtaken by newer RL-trained reasoning models
- · Maintenance is community-driven; no successor flagship from the same team in 2025-2026
- · Benchmark positioning is team-reported from its release moment