Price Per Token and OpenRouter list Prime Intellect’s INTELLECT-3 at $0.20 per million input tokens and $1.10 per million output tokens, with a 131K-token context window. Prime Intellect describes the model as a 100B+ mixture-of-experts release trained with large-scale reinforcement learning and published with open...
Price Per Token has listed Prime Intellect’s INTELLECT-3 API pricing at $0.20 per million input tokens and $1.10 per million output tokens.
Price Per Token’s model page, updated June 21, 2026, lists INTELLECT-3 with pricing of $0.200 per million input tokens and $1.10 per million output tokens. The same listing says the model has a 131K-token context window and is available through Prime-intellect API access.
OpenRouter’s Prime Intellect page separately lists INTELLECT-3 with the same 131K context window and the same pricing: $0.20 per million input tokens and $1.10 per million output tokens. OpenRouter also lists Prime Intellect model ID coverage on its page.
For developers, those figures provide a concrete basis for estimating usage costs before testing the model. The gap between input and output pricing also matters for application design. Workloads that produce long generated answers, reports, or code may be more affected by output-token costs than workloads that mostly analyze large inputs and return short responses.
Prime Intellect’s launch post describes INTELLECT-3 as a 100B+ mixture-of-experts model trained with large-scale reinforcement learning on the company’s RL stack. Prime Intellect says it released open-sourced weights, frameworks, datasets, environments, and evaluations alongside the model.
That release framing may interest researchers and developers who want to inspect or adapt parts of the model ecosystem. However, the practical ability to reproduce results still depends on factors such as available compute, implementation details, data access, and evaluation setup.
Price Per Token’s INTELLECT-3 page reports benchmark scores including 77.7 on LiveCodeBench, 82.2 on MMLU Pro, and 76.1 on GPQA. The source presents these as model-page benchmark figures, not as a full independent evaluation report.
Those numbers can help teams decide whether INTELLECT-3 is worth adding to an evaluation set, especially for coding, general knowledge, and difficult question-answering tasks. They should not be treated as a substitute for application-specific testing. Benchmarks do not fully capture latency, reliability, prompt sensitivity, tool-use behavior, safety constraints, or performance on proprietary domain data.
The most concrete public details in the available sources are the published token prices, the 131K context window, OpenRouter availability, and Prime Intellect’s description of the model release. Those details make INTELLECT-3 straightforward to compare on paper against other hosted large language models.
For production use, teams would still need to run their own tests. Useful checks include total cost under expected input and output lengths, response quality on real tasks, latency under load, handling of long-context prompts, and consistency across repeated calls.
Price Per Token and OpenRouter both list INTELLECT-3 at $0.20 per million input tokens and $1.10 per million output tokens, with a 131K-token context window. Prime Intellect describes the model as a 100B+ MoE system trained with large-scale reinforcement learning and released with open weights and related resources.
Price Per Token has listed Prime Intellect’s INTELLECT 3 API pricing at $0.20 per million input tokens and $1.10 per million output tokens.
Pricing and context window Price Per Token’s model page, updated June 21, 2026, lists INTELLECT 3 with pricing of $0.200 per million input tokens and $1.10 per million output tokens.
The same listing says the model has a 131K token context window and is available through Prime intellect API access.
Continue reading
