AICoder

AI Coding Live News

First-hand AI coding updates from 𝕏/Twitter frontier labs and researchers, distilled by AI with entity benchmarks.

Explore live on 𝕏 ↗

AI Coding Live News

X
Xuan-Son Nguyen / ggml-org@ngxson·
👀187
🔥96

llama.cpp ships Decision Models: /v1/systemone scores options in one forward pass (Julia-1, Laya, Kev-4B, lev, OpenJev)

On 2026-10-02 ggml-org announced llama.cpp server support for Decision Models via /v1/systemone (PR #29818, merged 2026-10-02T09:56Z). Send a state plus typed questions; get option probabilities in one forward pass. Day-0 models: Julia-1, Laya, Kev-4B, lev, OpenJev (vision). Cloudflare Clef is planned next.

llama.cpp ships Decision Models: /v1/systemone scores options in one forward pass (Julia-1, Laya, Kev-4B, lev, OpenJev)

Key Takeaways

  • Official post 2026-10-02 + merged PR #29818 (2026-10-02T09:56Z, +2139/−15)
  • Endpoint POST /v1/systemone with choice/score/noul; probabilities + confidence; output_tokens=0
  • Day-0 GGUFs: Julia-1 / Laya / Kev-4B / lev / OpenJev; median ~3–43 ms/question on RTX PRO 6000 (blog table)
  • Distinct from Ollama SystemOne & Cloudflare Clef hosting; Clef listed as next in llama.cpp
  • Start: llama serve -hf ggml-org/Kev-4B-GGUF then curl /v1/systemone
Read full story
Linked Models & Agents:
⚡ Julia 1⚡ laya⚡ Kev 4b⚡ lev⚡ openjev🌐 ggml-org🌐 huggingface
#llama.cpp#Decision Models#System One#Systemone#Jev#GGUF#local inference#agent routing#Kev-4B#OpenJev#Ggml
C
Comet@comet-ml·
👀54
🔥91

Comet Opik 2.2.84: online eval rules call Jev via OpenRouter Decisions API; queue automation + provenance

Comet Opik 2.2.84 (2026-09-29) lets online LLM-as-a-Judge rules call TypeSafe Jev end-to-end via OpenRouter POST /api/alpha/decisions (not chat/completions). Each Boolean metric becomes a yes/no question answered in one call; score 1 if probability ≥0.5, with Probability: 0.93 in the reason. Backend adds OpenRouterDecisionsClient (workspace OpenRouter key, 429/5xx retries) for ~typesafe/jev-latest / typesafe/jev-1.13. Also: Annotation Queue automation + item provenance. ~8 commits / 104 files vs 2.2.83.

Comet Opik 2.2.84: online eval rules call Jev via OpenRouter Decisions API; queue automation + provenance

Key Takeaways

  • Release 2.2.84; ~8 commits / 104 files vs 2.2.83
  • PR #8540: online judges call Jev via OpenRouter Decisions API
  • Boolean metrics → yes/no questions; score 1 if p≥0.5; reason shows Probability
  • Models: ~typesafe/jev-latest, typesafe/jev-1.13
  • Queue automation + item provenance; trace stats / thread_id fixes
Read full story
Linked Models & Agents:
⚡ Jev⚡ typesafe/jev-1.13🤖 LLM-as-a-Judge🤖 online evaluation🌐 Comet🌐 OpenRouter🌐 TypeSafe
#Opik#Jev#Openrouter#Decisions API#Evaluation#Release
O
Ollama@ollama·
👀141
🔥93

Ollama 0.35.0: Decision Models via /v1/systemone (Jev); Nimble & Tev1 return choices, probabilities, confidence

Ollama shipped v0.35.0 on 2026-09-28 (GitHub prerelease): Decision Models via TypeSafe’s Jev API at /v1/systemone. Instead of free text, models answer choice/noul/score questions with options, probabilities, and confidence—for ticket triage, model routing, and classification. First models: Nimble (Bespoke Labs) and Tev1 (Together AI). ~7 commits / 66 files vs v0.34.4; Settings no longer blocks on discovery; macOS update menu and stalled MLX downloads fixed.

Ollama 0.35.0: Decision Models via /v1/systemone (Jev); Nimble & Tev1 return choices, probabilities, confidence

Key Takeaways

  • v0.35.0 (prerelease) adds Decision Models at POST /v1/systemone (TypeSafe Jev)
  • Question types: choice (with probabilities), noul, score — not free-text generation
  • First models: Nimble (Bespoke Labs) and Tev1 (Together AI) on ollama.com/library
  • ~7 commits / 66 files vs v0.34.4; Settings/macOS update/MLX download fixes
Read full story
Linked Models & Agents:
⚡ Nimble⚡ Tev1⚡ Jev🤖 Decision Models🤖 model routing🌐 ollama🌐 TypeSafe🌐 Bespoke Labs🌐 Together AI
#Ollama#Decision Models#Systemone#Jev#Nimble#Tev1#Classification#Routing
A
Assaf Elovic@assafelovic·
👀83
🔥91

GPT Researcher 0.16/v3.7.0: Jev usefulness context filter beats embeddings by 59% relevance

GPT Researcher shipped v3.7.0 on 2026-09-26 (PyPI gpt-researcher 0.16.x). The context filter now defaults to TypeSafe Jev, scoring passage usefulness for each sub-query instead of embedding similarity. On a 28-task replay with identical sources, kept-passage relevance is 73% for Jev vs 46% for embeddings (~59% relative lift). Without TYPESAFE_API_KEY it falls back to local BM25—no embedding provider required. Also requires Python 3.12+ and adds a retriever plugin entry point.

GPT Researcher 0.16/v3.7.0: Jev usefulness context filter beats embeddings by 59% relevance

Key Takeaways

  • Shipped: v3.7.0 / PyPI 0.16.1 — pip install -U gpt-researcher
  • Official 28-task replay: Jev kept-passage relevance 73% vs embeddings 46% (~+59%) at similar cost
  • CONTEXT_FILTER=auto uses Jev with TYPESAFE_API_KEY else BM25; Jev failures fall back to keyword
  • Chunking: 50k page cap, 1k/100 overlap; keep ≤10 chunks (Jev/embeddings) or ≤25 (keyword) per sub-query
  • Requires Python 3.12+; retrievers via gpt_researcher.retrievers entry points
Read full story
Linked Models & Agents:
⚡ Jev⚡ TypeSafe-System-One🤖 GPT-Researcher🤖 Deep-Research-Agent🌐 GPT-Researcher🌐 TypeSafe
#GPT-Researcher#Jev#TypeSafe#Context filter#BM25#RAG#Deep-Research#Python-3.12
ADSponsored
C
Confident AI@confident-ai·
👀74
🔥90

DeepEval 4.2: Jev for LLM-as-judge decisions at ~$0.042/M input tokens and ~100ms

Confident AI’s deepeval 4.2.x (notes for 4.2.2/4.2.4; PyPI at 4.2.6) wires TypeSafe Jev into LLM-as-a-judge: evaluation LLMs still handle language tasks, while closed verdicts go to System One for typed decisions with calibrated probabilities. Claimed decision-stage pricing is about $0.042 per million input tokens with no output charge, ~100ms typical latency, and batched independent questions. Faithfulness, GEval, DAGMetric, and Classifiers are integrated; enable via deepeval set-mode experimental plus typesafe-sdk—missing keys fail loudly instead of silent fallback.

DeepEval 4.2: Jev for LLM-as-judge decisions at ~$0.042/M input tokens and ~100ms

Key Takeaways

  • Notes: python-v4.2.4; blog Introducing Jev for Evals; PyPI deepeval 4.2.6
  • Official: ~$0.042/M input tokens for decisions, no output charge; ~100ms typical; batch independent questions
  • Faithfulness/GEval/DAGMetric keep language work on the judge LLM; closed verdicts use Jev Noul/Choice/Score
  • New Classifiers for categorical labels (e.g. RefusalClassifier)
  • Opt-in: deepeval set-mode experimental + typesafe-sdk + TYPESAFE_API_KEY (loud failure if missing)
Read full story
Linked Models & Agents:
⚡ Jev⚡ TypeSafe-System-One🤖 DeepEval🤖 LLM-as-Judge🌐 Confident-AI🌐 TypeSafe🌐 DeepEval
#DeepEval#Jev#TypeSafe#LLM-as-judge#Faithfulness#GEval#Classifiers#Evals#CI
B
BerriAI@Berri_AI·
👀128
🔥94

LiteLLM 1.102.1: Claude Code safeguards on stable + TypeSafe Jev router/guardrails backported

BerriAI shipped LiteLLM v1.102.1 on 2026-09-23 (PyPI litellm 1.102.1). Versus 1.102.0 it backports Anthropic/Bedrock/Vertex safeguards→safeguard_results passthrough so Claude Code Auto mode through a LiteLLM gateway can show Auto mode server: Enabled, and cherry-picks seven TypeSafe Jev changes (passthrough, complexity router, relevance compaction guardrail, auto-router classifier, OpenRouter jev-1.13 pricing, etc.) from the 1.103 RC onto stable/1.102.x. Docker images keep cosign verify instructions.

LiteLLM 1.102.1: Claude Code safeguards on stable + TypeSafe Jev router/guardrails backported

Key Takeaways

  • Shipped: v1.102.1 / PyPI 1.102.1
  • Claude Code: gateway forwards safeguards and returns safeguard_results (#42538) so Auto mode server can stay Enabled
  • TypeSafe Jev: seven PRs backported to stable/1.102.x (#42595)
  • Deploy: pip install litellm==1.102.1; cosign-verify ghcr.io/berriai/litellm:v1.102.1
  • Docs: Proxy Quick Start + Anthropic provider pages
Read full story
Linked Models & Agents:
⚡ Claude⚡ TypeSafe Jev⚡ jev-1.13🤖 Claude Code🌐 berriai🌐 Anthropic🌐 openrouter
#LiteLLM#Claude Code#Safeguards#TypeSafe Jev#Proxy#Gateway#Cosign
A
AG2@ag2ai·
👀71
🔥93

AG2 1.1.0: TypeSafe Jev decision-only agents + reserved context strip across transports

AG2 v1.1.0 (2026-09-24) adds TypeSafe Jev decision-only agents via TypeSafeConfig + response_schema, and strips reserved ag:/a2a: context keys on A2A/AG-UI/A2UI/NLIP to block remote pre-approval of gated tools.

AG2 1.1.0: TypeSafe Jev decision-only agents + reserved context strip across transports

Key Takeaways

  • Decision agents: TypeSafeConfig + response_schema; bool/Enum/IntEnum as question shapes; probs on ModelMessage.metadata
  • Jev rejects tools and streaming; invalid schemas raise locally before the wire
  • Install: pip install "ag2[typesafe]"; see TypeSafe Jev blog + docs.ag2.ai
  • Security: strip reserved ag:/a2a: prefixes both ways on A2A/AG-UI/A2UI/NLIP
  • Also: MessageEnqueued, MCP required-arg validation + JSON-RPC codes, event retarget fix
Read full story
Linked Models & Agents:
⚡ Jev⚡ TypeSafe🤖 AG2🌐 AG2🌐 TypeSafe
#AG2#AutoGen#TypeSafe#Jev#decision agents#Security#MCP#Python
O
OpenClaw@openclaw·
👀121
🔥92

OpenClaw 2026.9.6: managed updates, restart recovery, 30-day Usage, plus Opus 5.5 / GPT-6 Sol·Luna / Grok 4.7

OpenClaw shipped v2026.9.6 on 2026-09-23 (macOS DMG rebuilt 2026-09-24 after launch crash #156861): clearer managed-update outcomes, restart recovery for unfinished work, complete 30-day Usage reporting, a GitHub reader, remote workspace Files/Memory/Skills, live meeting notes, optional TypeSafe Jev Decision Models, and chat models Claude Opus 5.5, GPT-6 Sol/Luna, and Grok 4.7. Scale: 2,614 PRs, 178 direct commits, ~350 contributors.

OpenClaw 2026.9.6: managed updates, restart recovery, 30-day Usage, plus Opus 5.5 / GPT-6 Sol·Luna / Grok 4.7

Key Takeaways

  • Verified scale on docs: 2,614 PRs / 178 commits / ~350 contributors
  • Managed updates + restart recovery; Usage covers a full 30-day window
  • GitHub reader; remote workspaces gain Files/Memory/Skills; live meeting notes
  • Decision Models (TypeSafe Jev/local) + Opus 5.5, GPT-6 Sol/Luna, Grok 4.7
  • Install [email protected]; macOS users hit by #156861 should reinstall the rebuilt DMG (#156881)
Read full story
Linked Models & Agents:
⚡ Claude Opus 5.5⚡ GPT-6 Sol⚡ GPT-6 Luna⚡ Grok 4.7⚡ TypeSafe Jev🤖 OpenClaw🌐 OpenClaw
#OpenClaw#Agent#Usage#restart recovery#Opus 5.5#GPT-6#Grok 4.7
s
sora19ai@sora19ai·
👀60
🔥89
View on 𝕏 ↗

jev-mcp open-sources nine TypeSafe Jev judgment tools for Cursor and Codex

@sora19ai highlighted burnigtm/jev-mcp, a local stdio MCP that exposes TypeSafe Jev structured judgments to Cursor, Codex, and other MCP clients while the host still edits files and runs commands. Nine tools include jev_step, jev_review, jev_verify, and jev_rank for patch review, claim checks, and ranking; a TypeSafe API key is required, with mock mode for demos. MIT-licensed with about 13 GitHub stars at verification.

Key Takeaways

  • Local MCP that puts Jev judgments on Cursor/Codex loops while the host still executes tools.
  • Nine tools cover step routing, patch review, claim verification, screening, and ranking.
  • Needs a TypeSafe API key; burnigtm/jev-mcp under MIT with about 13 stars at verification.
Read full story
Linked Models & Agents:
🤖 Claude Code (Anthropic)🤖 Cursor Composer (Agent Mode)🤖 Aider🌐 typesafe
#Jev MCP#TypeSafe Jev#MCP#Cursor#Codex#patch review#open source
R
Roshan Chandna@roshanchandna·
👀40
🔥89
View on 𝕏 ↗

Auto-Guard open-sources Jev-backed allow/escalate/block gates for coding-agent tool calls

@roshanchandna introduced rchandnaWUSTL/auto-guard: before a coding agent runs a tool call, Auto-Guard sends the proposal to TypeSafe Jev and returns allow, escalate (ask you with a one-line LLM reason), or block. Install with pip install agent-autoguard then autoguard install for Claude Code by default, with adapters for Cursor, Codex, Copilot, Gemini, and Windsurf. The author cites roughly 0.4s and about $0.000025 per check.

Key Takeaways

  • Pre-tool-call gate with Jev verdicts: allow, escalate for human approval, or block.
  • pip install agent-autoguard then autoguard install for Claude Code; adapters cover Cursor, Codex, and peers.
  • Author cites ~0.4s and ~$0.000025 per check; rchandnaWUSTL/auto-guard publishes code and known limits.
Read full story
Linked Models & Agents:
🤖 Claude Code (Anthropic)🤖 Cursor Composer (Agent Mode)🤖 Aider
#Auto-Guard#Agent autoguard#Jev#TypeSafe#Claude Code#tool calls#Safety#open source
ADSponsored
h
harsh@harsh_w98·
👀56
🔥87
View on 𝕏 ↗

Open-source Jev Sentinel: TypeSafe Jev guards what the Pi coding agent does, reads, and says

@harsh_w98 released open-source extension harshwasan/pi-jev-sentinel for the Pi coding agent: TypeSafe Jev checks what the agent does, reads, and says; .env files are withheld from Jev and quoted secrets scrubbed (with a caveat that env/grep paths can still leak), plus code and test sandboxes.

Key Takeaways

  • Pi coding-agent extension: Jev reviews actions, reads, and speech.
  • .env withheld from Jev with scrubbed quotes; code/test sandboxes included.
  • Open-source harshwasan/pi-jev-sentinel on TypeSafe Jev.
Read full story
Linked Models & Agents:
#Jev Sentinel#Jev#Pi#coding agent#TypeSafe#Safety#open source
Ά
Άντιpacman🪽@doth4580·
👀50
🔥88
View on 𝕏 ↗

mcp-typesafe: Rust MCP server exposing TypeSafe Jev System One tools to any agent

Community release mcp-typesafe (github.com/sf-stav/mcp_typesafe): an unofficial Rust MCP server that exposes TypeSafe System One / Jev structured judgments (choice/score) plus higher-level filter/rank/classify/choose-one and stability tools to any MCP client. It targets bulk triage without burning LLM context and calibrated probabilities instead of prose confidence; most calls are ~100ms. stdio and SSE transports, TypeSafe API key required, single static binary.

Key Takeaways

  • Unofficial Rust MCP exposing Jev/System One structured judgments to any agent.
  • Bulk filter/rank/classify/choose-one plus stability sampling; most calls ~100ms.
  • stdio + SSE; TypeSafe API key required; single static binary.
Read full story
Linked Models & Agents:
⚡ jev🌐 typesafe
#MCP typesafe#MCP#Jev#TypeSafe#System One#Rust#decision model
G
GitHub Next@GitHubNext·
👀140
🔥91
View on 𝕏 ↗

GitHub Next opens LocalJev: local Jev-compatible decision API on oMLX

GitHub Next open-sourced LocalJev (github.com/githubnext/localjev): a morning build of a “poor man’s Jev” for teammates without TypeSafe Jev access. It exposes a Jev-compatible POST /v1/systemone API in Bun/TypeScript, defaults to oMLX-backed local models such as DiffusionGemma, and reports evals across DiffusionGemma, Qwen MoE, Gemma 4 MoE, and Gemma 4 e4b/e2b (benchmark report in-repo). The LocalJev server works with normal Jev client wrappers, but probabilities are model-reported rather than logit reads—calibrate before consequential use.

Key Takeaways

  • Open-sources LocalJev: a local Jev-compatible POST /v1/systemone decision API (Bun/TypeScript).
  • Defaults to oMLX-backed local models such as DiffusionGemma, with multi-model evals in-repo.
  • Works with normal Jev client wrappers; probabilities are model-reported, not logit reads—calibrate first.
Read full story
Linked Models & Agents:
⚡ diffusiongemma⚡ qwen⚡ Gemma 4🌐 github
#LocalJev#GitHub Next#Jev#TypeSafe#oMLX#DiffusionGemma#local LLM#decision model
T
Taro L. Saito@taroleo·
👀71
🔥92
View on 𝕏 ↗

Dev distills DeepSeek V4 Flash decisions into a 4B local model: 26h on DGX Spark, ~22ms per call

Developer Taro L. Saito (@taroleo) followed a TypeSafe Jev-style approach—dropping full text/JSON generation and specializing on decisions—to distill DeepSeek V4 Flash judgments (~157GB weights) into a 4B model with about 26 hours of training on a DGX Spark. The student is roughly 1/20 the size, beats the teacher’s instant-answer mode, and runs about 22ms per decision, showing a fast local-LLM decision engine is practical.

Key Takeaways

  • Jev-style decision-only specialization distills DeepSeek V4 Flash judgments into a 4B model.
  • About 26 hours on a DGX Spark; roughly 1/20 the teacher size and ~22ms per decision.
  • Claims to beat the teacher's instant-answer mode—useful for local routing/gating agents.
Read full story
Linked Models & Agents:
⚡ DeepSeek-V4-Flash🌐 DeepSeek
#DeepSeek V4 Flash#Distillation#Jev#DGX Spark#local LLM#decision model#4B
V
Vercel@vercel·
👀66
🔥91
View on 𝕏 ↗

Vercel: TypeSafe Jev is the fastest-adopted model in AI Gateway history

Vercel reports TypeSafe Jev became the fastest-adopted model in AI Gateway history within 24 hours of launch: nearly 13% of paid teams on day one, about 2x the GPT-5.6 family and over 6x Fable 5.1. It crossed 10% of teams in about 18 hours, while other recent launches stayed below 7% after a full day. The post frames durability of that early lead as the next test.

Key Takeaways

  • Official: Jev reached ~13% of paid AI Gateway teams on day one, the fastest adoption on record.
  • About 2x GPT-5.6 family and 6x+ Fable 5.1; ~10% of teams within 18 hours.
  • Details at vercel.com/blog/ai-gateway-jev-model-launch; next question is whether the lead lasts.
Read full story
Linked Models & Agents:
⚡ jev🌐 vercel🌐 typesafe-ai
#Vercel#AI Gateway#Jev#TypeSafe#Adoption#GPT-5.6#Fable 5.1