OmniCoder-9B v2 drops with agentic stability

// 127d agoMODEL RELEASE

OmniCoder-9B v2 drops with agentic stability

Tesslate releases OmniCoder-9B v2, a ground-up rebuild of the 9B coding model optimized for agentic reasoning and long-context stability. The model addresses core repetition issues from v1 while achieving frontier-class performance on reasoning and terminal-based benchmarks.

// ANALYSIS

OmniCoder-9B v2's shift to assistant-only training and a constant learning rate schedule marks a significant leap in model reliability for autonomous coding loops. The version achieves near-zero repetition loops in multi-turn tool calling tasks and inherits a 262K context window from Qwen3.5-9B. It outperforms significantly larger models on Terminal-Bench 2.0 while maintaining high efficiency for local execution on mid-tier consumer hardware.

// TAGS

omnicoder-9bllmai-codingagentopen-weights

DISCOVERED

127d ago

2026-03-25

PUBLISHED

127d ago

2026-03-24

RELEVANCE

9/ 10

AUTHOR

Western-Cod-3486

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

UPDATE18m ago

ElevenLabs introduces Character Casting for Audiobooks

ElevenLabs has introduced Character Casting for Audiobooks, a feature that automatically analyzes uploaded manuscripts to detect individual characters and propose matching AI voices. Creators can preview proposed voices reading actual dialogue directly from their manuscript and apply global pronunciation rules across chapters.

UPDATE41m ago

B.AI launches API resource group with 10% discount

B.AI (TheB.AI) announced the release of its official API Resource Group, featuring a 10% discount designed to support developers, startups, and enterprise users. As AI integration accelerates, B.AI aims to address the demand for accessible, high-performance, and cost-managed AI infrastructure by streamlining API access and reducing overall operational expenses.

BENCHMARK44m ago

Merge Gateway cuts LLM costs 65%

Merge released benchmark data showing intelligent model routing cuts average task costs by 65% ($2.87 vs $8.17) while preserving 99.6% accuracy compared to fixed Opus 4.8. Routing overhead remained minimal with a median latency of 90–650ms per request across 120 trials.