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

A
AgentField / Santosh Radha@Agent-Field·
👀82
🔥95

AgentField open-sources CodeAF v0.6: one Go binary software factory for open models; vendor DeepSWE 54.9% #1 same-model

CodeAF is an Apache-2.0 Go binary that runs open-weight models as a local software factory with task fan-out and Pareto crewing. v0.6.0 shipped 2026-10-01; ★259 now. Vendor DeepSWE doc: /senior-dev solved 62/113 (54.9%) on DeepSeek V4 Flash vs nine other harnesses—with stated one-seed and sampling caveats.

AgentField open-sources CodeAF v0.6: one Go binary software factory for open models; vendor DeepSWE 54.9% #1 same-model

Key Takeaways

  • Agent-Field/CodeAF Apache-2.0; ★259 / 34 forks; install via agentfield.ai/get/codeaf
  • v0.6.0 published 2026-10-01T18:59:00Z; README still says early preview
  • Vendor DeepSWE same-model table: senior-dev 62/113 (54.9%) at ~22¢/task vs nine harnesses—one seed; sampling-contract caveat documented
  • Later same 113 tasks: 88/113 (77.9%) on DeepSeek V4.1 Flash; 78/113 (69.0%) on Kimi K3 (senior-dev only)
  • Built-in providers include OpenRouter, DeepSeek, GLM, Kimi, MiniMax, Qwen, Codex, Ollama; default worker/planner/checker crew
Read full story
Linked Models & Agents:
⚡ DeepSeek-V4-Flash⚡ DeepSeek-V4.1-Flash⚡ Kimi K3🤖 CodeAF🤖 senior-dev🌐 AgentField🌐 DeepSeek🌐 Moonshot
#Codeaf#Agentfield#Open models#Deepswe#Coding harness#Software factory
T
Taro L. Saito@taroleo·
👀67
🔥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
K
K2S@k2sbhai·
👀45
🔥86
View on 𝕏 ↗

Fable Orchestrator Codex skill routes work to Luna, DeepSeek V4 Flash, and Fable 5.1

Open-source Codex skill Fable Orchestrator (k2s1/Fable-Orechester-) routes one task across specialized models: GPT-5.6 Luna for implementation, DeepSeek V4 Flash for loops, and Claude Fable 5.1 for planning and final adjudication. After install, run /fable orchestrator [task].

Key Takeaways

  • Shape: Codex skill at k2s1/Fable-Orechester-.
  • Routing: Luna implement, DeepSeek V4 Flash loops, Fable 5.1 plan/adjudicate.
  • Usage: install then /fable orchestrator [task].
Read full story
Linked Models & Agents:
🤖 Claude Code (Anthropic)
#Fable Orchestrator#Codex#Fable 5.1#GPT-5.6 Luna#DeepSeek V4 Flash#Skills#Multi agent
F
Fireworks@FireworksAI_HQ·
👀63
🔥88
View on 𝕏 ↗

Fireworks ships DeepSeek V4 Flash with native vision at $0.22/$0.66 per 1M tokens

Fireworks said DeepSeek V4 Flash now includes native vision on serverless and priority tiers. It reports better benchmarks than the text-only 0731 checkpoint at the same $0.22 input / $0.66 output per 1M tokens. The model page lists a 305B MoE with about 1040K context, function calling, image input, and a speculative-decoding module.

Key Takeaways

  • DeepSeek-V4-Flash-Vision-Exp is live on Fireworks serverless and priority tiers with native image input.
  • Same price as text-only 0731: $0.22 input / $0.66 output per 1M tokens (cached input $0.007).
  • 305B MoE, ~1040K context, function calling, and speculative decoding for screenshot-aware coding agents.
Read full story
Linked Models & Agents:
⚡ DeepSeek-V4 (Next-Gen MoE)🤖 Cline (VS Code Agent)🌐 DeepSeek
#DeepSeek-V4-Flash-Vision#Fireworks#Serverless#Pricing#Multimodal#Agentic coding
ADSponsored
C
Composio@composio·
👀62
🔥89
View on 𝕏 ↗

Composio Evaluates 5 Open-Weight Models on 30 Multi-Step Agent Tasks: Flash Models Lead in ROI

Composio benchmarked 5 open-weight models across 30 multi-step agent tasks, finding GLM-5.3 Flash and DeepSeek-V4 Flash delivered the highest cost efficiency at a fraction of full-model pricing.

Key Takeaways

  • Flash-class models achieve near parity with full-scale weights at approximately 75% lower inference cost;
  • Multi-application tool orchestration remains the primary failure mode across all open-source models;
  • Provides empirical cost-per-success metrics to guide architectural selection for enterprise agent deployments.
Read full story
Linked Models & Agents:
⚡ GLM-5.3-Flash⚡ DeepSeek-V4-Flash
You've caught up with all stories