PRICES
MINIMAX-M2$1.001000K
QWEN3-CODER-FL$0.981000K
MINIMAX-M2.5$0.991000K
GROK-4-1-FAST-$0.502000K
GROK-4-FAST$0.502000K
GEMINI-3.1-FLA$1.501049K
GROK-4-FAST-NO$0.502000K
QWEN-PLUS-2025$0.781000K
QWEN-PLUS$0.781000K
GROK-4-FAST-RE$0.502000K
GPT-4.1-NANO$0.401048K
GPT-4.1-MINI-2$1.601048K
GPT-4.1-MINI$1.601048K
MINIMAX-M2.7$1.201000K
GROK-4.1-FAST$0.502000K
LLAMA-4-MAVERI$0.601049K
MINIMAX-01$1.101000K
QWEN-PLUS-2025$0.781000K
DEEPSEEK$0.281049K
MINIMAX-M3$1.20Elo 1528
GEMINI-2.5-FLA$0.401049K
GROK-4-1-FAST-$0.502000K
GEMINI-2.5-FLA$0.401049K
GEMINI-2.0-FLA$0.301049K
GEMINI-2.0-FLA$0.401049K
MINIMAX-M2.1$0.951000K
QWEN3.5-FLASH-$0.261000K
DEEPSEEK-V4-FL$0.281049K
MINIMAX-M2$1.001000K
QWEN3-CODER-FL$0.981000K
MINIMAX-M2.5$0.991000K
GROK-4-1-FAST-$0.502000K
GROK-4-FAST$0.502000K
GEMINI-3.1-FLA$1.501049K
GROK-4-FAST-NO$0.502000K
QWEN-PLUS-2025$0.781000K
QWEN-PLUS$0.781000K
GROK-4-FAST-RE$0.502000K
GPT-4.1-NANO$0.401048K
GPT-4.1-MINI-2$1.601048K
GPT-4.1-MINI$1.601048K
MINIMAX-M2.7$1.201000K
GROK-4.1-FAST$0.502000K
LLAMA-4-MAVERI$0.601049K
MINIMAX-01$1.101000K
QWEN-PLUS-2025$0.781000K
DEEPSEEK$0.281049K
MINIMAX-M3$1.20Elo 1528
GEMINI-2.5-FLA$0.401049K
GROK-4-1-FAST-$0.502000K
GEMINI-2.5-FLA$0.401049K
GEMINI-2.0-FLA$0.301049K
GEMINI-2.0-FLA$0.401049K
MINIMAX-M2.1$0.951000K
QWEN3.5-FLASH-$0.261000K
DEEPSEEK-V4-FL$0.281049K
Model Launches· MarkTechPost· Jun 7, 2026· 17 hours ago· 1 min read

Meet Harness-1: A 20B Retrieval Subagent Trained With Reinforcement Learning Inside a Stateful Search Harness on gpt-oss-20b

UIUC and Chroma's Harness-1 is a 20B retrieval subagent trained with reinforcement learning inside a stateful search harness. The harness maintains the bookkeeping — candidate pool, importance-tagged curated set, eviden…

Why it matters

New models reset the capability and price-performance frontier. Teams re-evaluate what to build on whenever a launch shifts what's possible per dollar.

Explore the data behind this

Related HotON.ai pages

Models →Compare →
Read original (MarkTechPost) →

Summaries are aggregated for information only — follow the source link for the full story. Demo entries are illustrative.

More news

Model Launches4 hours ago
Mythograph Atelier #1 - Abstract Art That Means Something to You
Model Launches6 hours ago
Notion restores access to Anthropic after service disruption
Model Launches7 hours ago
Building Reflective Prompt Optimization with GEPA: Multi-Component Prompts, Structured Feedback, and Held-Out Validation
Model Launches7 hours ago
OpenAI is still working on that ‘super app’