Model Routing

OpenRouter launched Auto Code, reporting Opus-level coding performance at 74% lower cost. Its routers explain decisions and adapt to model restrictions; beta reportedly cut invalid tool calls fourfold.

Vals tested seven models and found four retained at least 80% of their scores at half the cost or less. All seven scored worse overall at their lowest reasoning setting.

Copilot Fusion coordinates Claude and GPT models for coding work. In a GitHub-funded Artificial Analysis test, Fusion Max cost $5.59 per task versus $11.42 for Claude Code.

Shreya Shankar describes a coding workflow built around human-reviewed RFCs, implementation routing by task size and separate bug review. Follow-ups explain section-by-section specs and restarting failed approaches.

An independent DeepSWE test found Jev Router roughly matched GPT-6 Astra low on results but cost slightly more and took nearly five times as long. Practitioners also argue that request-level routing misses repository context and cache costs.

OpenRouter launched Jev Router, which selects a model and reasoning effort per turn while weighing the cost of losing cached context. OpenRouter reports 237 of 423 tasks solved; invalid Jev outputs or timeouts fail without fallback.

Amp says its coding agent is free when users supply their own compute, model subscription, or API key. The rollout includes routing support for external providers such as Ollama Cloud, OpenRouter, and custom OpenAI-compatible URLs.

Fusion keeps a lead model in control while routing execution to a cheaper model in Devin CLI. Cognition reported 39% lower benchmark cost, and Artificial Analysis measured 43% lower cost with 31% faster runs at near-frontier scores.

Sakana says Fugu Max dynamically routes requests across open-weight and specialist models at two to six times lower cost than elite models. In the same release, the company says Fugu Ultra v2 led five of eight hard evaluation suites.

A benchmark author says OpenRouter likely routed GLM-5.3-Flash requests to quantized endpoints because precision was not pinned. Twelve reruns using pinned FP8 and self-hosted inference improved results, suggesting earlier scores may have reflected routing.

Posts say OpenAI will terminate Cursor's direct model access on November 12. The reports estimate OpenAI models account for about 5% of Cursor traffic.

Glean says its runtime routes enterprise-agent work across more than 40 models using company context. It reports $0.58 per query and 78% user preference over Claude Cowork in a 180-person benchmark.

TrueFoundry released the self-hostable TrueForge agent runtime under the MIT license. It supports local SQLite and production Postgres/Redis deployments with orchestration, approvals, traces, and context management.

OpenRouter upgraded Auto Router to classify prompts into about 30 task types, then route by anonymized 7-day spend share and cost tier. OpenRouter says the max tier beat the old router across five benchmark domains.

Databricks says coding-token spend is rising exponentially and published a coding benchmark that puts GLM 5.2, Claude Opus 4.8, and GPT-5.6 Sol on the quality-per-dollar frontier. Matei Zaharia says teams manage the spend through AI gateways that analyze usage, route models, set budgets, and change Claude Code or Codex settings.

Practitioners reported better Codex multi-agent runs by raising concurrency and splitting work across Sol, Terra, and Luna. One workflow sends deploy tasks to Luna Max to preserve Sol tokens.

Sol-advisor routes Codex tasks through GPT-5.6 Sol, Luna, and Terra, while users compare Luna Max as a lower-cost reasoning setting. Early reports say small routing tests need larger benchmarks.

Wafer listed Kimi K3 Fast on OpenRouter and Vercel AI Gateway. It claimed 172 output tokens/sec, 15.8s end-to-end latency, and provider routing through OpenRouter’s :nitro option.

Microsoft said MAI-Cyber-1-Flash inside the MDASH multi-agent security harness scored 95.95% on CyberGym. The system routes harder tasks to GPT-5.4 and coordinates more than 100 specialist agents.

Cursor launched Router for Teams and Enterprise with Intelligence, Balance, and Cost modes. Cursor says the router can select models per request and cut costs by 60%, with admin controls for businesses.

Martian launched Ship, a beta endpoint that sits between an app and its reference model and routes each request through cheaper paths while aiming to preserve behavior. Martian says the beta targets 50% lower cost.

OpenCode users reported complex DeepSeek V4 API prompts returning Fable-like outputs and CoT style, while simpler prompts did not. The evidence is community-led and disputed, so the claim remains unverified.

AI21 says its pipeline routes exploration, extraction, and patching across model tiers and reaches 80.8% on SWE-Bench Pro at $5.99 per task. Other workflows use critic agents, readable harnesses, Fugu/Nemotron routing, and scheduled Gemini agents.

Cognition said Devin Fusion now uses Fable 5 and saw lower cost per task than Opus 4.8. Practitioners cited Fable-led delegation patterns that cut token use, with caveats on serial debugging.

Devin exposed routers such as Fusion, claiming frontier performance at 35% lower cost, while Databricks argued for smart routing in agent harnesses. New charts put Grok 4.5 and Muse Spark 1.1 near the coding-cost frontier.

Databricks published an internal coding-agent benchmark using tasks from its codebase. OpenAI, Anthropic, and GLM-5.2 models landed on its Pareto frontier, and the company argues teams should optimize cost per task rather than per token.

ClaudeDevs reports Sonnet 5 with a Fable 5 advisor reached ~92% of Fable 5's SWE-bench Pro score at ~63% of the price. Other builders route implementation to Sonnet, Codex, GPT-5.5, or GLM workers.

OpenRouter published an MCP workflow that it says cut inference costs 24x at comparable quality. The MCP lets the model choose providers using codebase context plus OpenRouter benchmark, aggregate-usage, and live-performance data.

Codex app code now references GPT-5.6 Sol, Terra, and Luna, while posts claim Sol Ultra reaches 91.9% on TerminalBench at lower cost. Treat release timing, limits, and benchmark claims as unofficial until OpenAI publishes details.

Cognition launched Devin Fusion, a hybrid coding harness that reroutes work mid-task and says it cuts Fable-class cost by 35%. Use it when upfront routing misses late complexity; the router can re-evaluate after investigation starts.

OpenRouter said four open-weight models now handle real agentic workloads, and a JPMorgan report put Chinese models at about 45% of platform traffic. The shift matters because teams are optimizing for price, hosting, and task fit instead of defaulting to frontier APIs.

Sakana made Fugu Ultra available through Vercel AI Gateway, while new technical writeups described the trained routing head and multi-step orchestration behind it. The integration matters because teams can invoke Fugu’s model-selection workflow through existing gateway plumbing instead of standing up custom routing.

OpenRouter released an MCP server that lets agents query live model pricing, benchmark scores, provider data, docs, and run test inference from the CLI. That replaces stale model knowledge with current routing data inside long-running agent workflows.

OpenRouter released a dedicated Image API that normalizes request shapes across 30-plus models from eight providers. Agents can inspect limits, passthrough options, streaming, and exact per-call cost without hardcoding vendor quirks.

Kilo Code added an Auto Efficient mode that routes each request to the cheapest model that clears its benchmark bar using public KiloBench results. The router stays session-aware and falls back to stronger paid models when confidence is low.

Sakana launched Fugu Ultra on AI Gateway and published a technical report, with early testers sharing mixed results. Reports mention polished outputs on some tasks, but also 30-minute runs, uneven coding quality, and much higher cost than GLM-5.2.

Morph said its code-serving stack now exposes Qwen, GLM-5.2, MiniMax M3, and DeepSeek v4 with code-tuned speculative decoding. It claims 20-35% higher acceptance than Eagle 3.1 or DFlash, plus kernels for cheaper hardware.

Codex workflows can now run against open-weight models served through compatible Responses API endpoints, with Ollama and vLLM publishing direct paths for GLM-5.2 and Kimi K2.7 Code. That matters because teams can keep the Codex interface while swapping to self-hosted or lower-cost inference backends.

OpenRouter launched Fusion, a server-side panel API that sends prompts to multiple models and combines one answer. Early logs also showed a web-path issue where Fusion still invoked Claude Opus 4.8 as judge and billed for it until API-side control was clarified.

OpenRouter launched Fusion, a server-side panel API that fans prompts to multiple models, judges the outputs, and returns one synthesized answer. The company said DRACO landed within 1% of Fable at roughly half the price, but the published evals do not cover long-horizon tasks.

One day after Fable 5 launched, users reported burning through Max quotas in about 90 minutes while Anthropic told subscribers the model will leave Claude plans on June 23 until capacity improves. If you depend on Fable, plan for quota pressure and route critical jobs elsewhere.

OpenRouter, OpenCode, Lovable, Cline, Browser Use Terminal, Nous Portal, and Venice all added Fable 5 within hours of launch. The rollouts put the model into gateways, coding agents, browser agents, and chat clients on day one.

Vendors pushed routing and spend controls closer to the default app layer, including OpenRouter's cache-hit pricing telemetry and Devin's adaptive routing. The discussion frames model choice more as a budget-control problem than a pure quality setting.

OpenRouter launched Pareto Code, a free experimental coding router that filters by min_coding_score and says it is already handling about 1 billion tokens a day. The release adds a tunable routing path for coding workloads where cost and model quality need to be balanced.

Factory put Router into private preview in its CLI and desktop app to route coding tasks across models, claiming 20-25% lower spend. The launch targets rising agent costs, though session continuity and routing behavior remain active points of debate.

Perplexity said Computer will split tasks between on-device models and frontier cloud models, keeping some data on the local machine while escalating harder work remotely. That matters for privacy-sensitive workflows and for reducing token-heavy cloud usage on laptop-class hardware.

Independent developers shipped sidecars that let Claude Code, Cursor, and Codex share memory, hot-swap model providers, package local projects as apps, and automate browser QA. Try these reusable tools if you want memory, routing, QA automation, and app packaging outside editor-specific features.

OpenRouter released Guardrails to apply budget limits, provider restrictions, zero-data-retention rules, prompt-injection defense, and DLP checks across routed traffic. Google Model Armor and Lakera Guard connectors are in beta, so plan around limited availability.

Independent IDEs, gateways, and agent runtimes rolled out Claude Opus 4.8 within hours of launch, including Cursor, Warp, OpenRouter, and Perplexity. That matters because teams can benchmark or swap the model into existing workflows without waiting for connector lag.

Hermes Agent added a built-in MCP Catalog while separate builders shipped Qwen3.7 Max support, Venice private-model workflows, and Krea 2 image generation. The cluster shows Hermes moving beyond a single-model assistant toward a broader agent shell with tool, model, and media providers.