Pokee-Isaac 28B Launches: 10M-Token Context on a Single GPU
Kyrannio · x · 2026-08-05
Pokee AI announced Pokee-Isaac 28B, claiming to be the world's first real 10M-token context frontier-class agentic model deployable on a single GPU (starting from RTX 4090).
Featuring a new proprietary non-decoder-only architecture, the model boasts:
- 93.3% RULER at 10M tokens.
- Up to 137K tokens/s prefill on a single B200.
- Leading performance on BFCL v4 and τ³-bench.
- Lowest combined attack success rate on the DTAP security benchmark.
Priced aggressively at $0.15/M input and $1/M output, it supports on-prem, VPC, and on-device deployments with Day-0 support for vLLM and SGLang.
Related event: Pokee-Isaac 28B Launches: 10M Context on Single GPU, Tops BFCL v4(5 posts)→
More from coding & agent
- NeurIPS 2026 Call for Papers: Focus on Responsible Meta-Agents — RobertTLange · 2026-08-05
- Spring Evals: Benchmarking LLMs on Real-World Spring Boot 4 Code with Hidden Tests — therealdanvega · 2026-08-05
- Navigating SAP's Agent Architecture: MCP for Tool Execution, A2A for Orchestration — SillyDot3305 · 2026-08-05
- YC's Multiplayer Agent Harness: Great Decoupled Architecture, But It Kills KV Cache — Instance_Not_Found · 2026-08-05
- The 'oh-my' Configuration Trend in Open-Source Coding Agents — HarveenChadha · 2026-08-05
- Ex-Google Engineer Releases 3-Hour Free Course on Building AI Agents — ___Mufasaa · 2026-08-05