Excited to share that TwinRouterBench has been accepted to the #RLEval# Workshop at #CAIS2026# 🎉 As LLM apps become long-horizon agents, one request can trigger many model calls across planning, tool use, retrieval, coding, and verification. That makes per-step LLM routing a core infrastructure problem: sending each call to the cheapest sufficient model without breaking downstream success. TwinRouterBench introduces: ⚡ Static track: 970 router-visible prefixes from 520 instances across SWE-bench, BFCL, mtRAG, QMSum, and PinchBench 🚀 Dynamic track: live SWE-bench Verified evaluation with official task resolution + realized API spend Key result: a router trained on static labels achieves comparable SWE-bench resolve rate while cutting API cost by ~53% vs. an unrouted Opus 4.6 baseline. Paper: Code: Dataset: Website: #LLM# #AgenticAI# #LLMRouting# #Benchmark# #SWEBench#

TL;DR: No single LLM is optimal for every query and budget. This work unifies the scattered field of LLM routing into "five building blocks" and ships an open-source infrastructure bundling 16+ routers with a dedicated benchmark. Title: LLMRouter: Unified Infrastructure for Developing, Evaluating, and Deploying LLM Routers URL: Key points 🧩 Decomposes any router into 5 components: Context Encoder / Model Encoder / Scoring / Decision Rule / Learning Signal 📊 xRouteBench spans 5 tracks (Generic, Memory, Vision, TimeSeries, Personalized), 4,767 queries total 🛠️ A dense query-model matrix, built by dispatching every query to all 18 candidate models, serves as both training supervision and test bed 🚀 Learned routers beat the strongest fixed-model baseline (always pick the largest) by 14.6% relative ⚖️ No router dominates: RouterDC leads on accuracy but drops to 10th under tight cost constraints 🔁 Multi-turn routing adds cost without consistent gains (Router-R1 at just 22.3%) 👤 User conditioning helps, but the best design flips between simulated personas and real feedback A solid step that proves "the best router depends on task and budget" and lays a foundation for fair comparison. #LLM# #ModelRouting#

Omnigent smarter routing picks harness + model from your task. 🔀 Auto · smart routing harness option ⚙️ Activates automatically from llm:/routing: config (no OMNIGENT_SMART_ROUTING env var to set anymore) 🎛️ Composer gear modal: tune model, reasoning effort, and routing mid-chat Learn more: #Omnigent# #AIAgents#

Daily token throughput hit 1.51 trillion—a 71,500x+ surge in 5 months. is rapidly scaling as the next-gen AI infrastructure. From LLM routing and compute scheduling to Agent execution and automated settlement, sits "Above all models, beneath all Agents"—building the global intelligence settlement layer for the Agentic era. Powered by Web2 + Web3 payment rails, a full-stack Agent execution environment, and protocols like x402 and 8004, seamlessly connects compute, tasks, and capital. 📖 Read the deep dive:

TL;DR Routing between models at the token level looks great on paper, yet running it on today's serving stacks makes things slower. This paper fixes that with a purpose-built runtime. Accepted at NeurIPS 2026. Title: TokenRouter: Efficient Serving System for Token-Level LLM Routing URL: Points 🎯 R2R matches 32B quality while sending just 5% of tokens to the 32B model, versus roughly 40% of queries under query-level routing 🧱 But vLLM and SGLang assume every request is step-synchronous, so fast models idle waiting on slow ones and batches fragment on every switch ✍️ Developers write only route, send, and receive, tracing how a single request flows, while the runtime handles async execution 🔀 Each model gets its own subserver, with client, decoding, and inter-model loops decoupled and KV-cache preserved across handoffs ⏳ Instead of firing on arrival, the scheduler buffers to a threshold; wait too long and the small model starves, so the optimum is derived from a discrete-time Markov chain 📈 Wins on all 15 algorithm-by-workload combinations, with 2.01-64.15x higher throughput 🧪 Throughput holds steady when output grows from 2,048 to 8,192 tokens, where the standard baseline loses 58-85% of its own The range is wide though, and on Co-LLM and R-Stitch the 32B model alone still runs faster. Worth remembering this isn't a universal win. #LLM# #Inference#

You may have heard that Legora has surpassed $200M in ARR at positive gross margins. Here is the other half of the story: what happens after customers are on the platform. Usage has to map to value, and efficiency gains must show up for customers. Most credit consumption on @WeAreLegora comes from two areas: LLM usage and document processing. In August, we significantly improved our Agent harness, LLM routing, and orchestration. Credits tied to LLM usage on Agent dropped by roughly 25%, with response quality improving. In September, we rebuilt how Legora ingests and processes documents. The work now uses ~25% fewer credits while maintaining quality. Those two changes have reduced credit consumption for the same work by more than 20% since August. Consumption-based customers can now do the same work for fewer credits. Next, we are working on additional harness and routing improvements, as well as more efficient document processing. As we find efficiency gains, our goal is to keep passing them back to consumption-based customers. Legora is building a long-term company to support legal teams for decades to come. Our strong financial health makes that all possible. That includes a pricing model that gives back to our customers.