AI Coding Live News

GitHub staffer Pierce Boggan (@pierceboggan) announced Copilot Auto model selection now offers three tiers—efficiency (cost-first), balance (cost/quality/latency), and intelligence (quality-first). Auto still picks a model per prompt; usage is billed on the selected model, and paid subscribers keep a 10% Auto usage discount. Rolling out in VS Code, Copilot CLI, and the GitHub Copilot app (changelog 2026-09-14).

Artificial Analysis (@ArtificialAnlys) reports Google DeepMind’s Gemini 3.8 Flash (high reasoning) scores 59 on the Intelligence Index, +3 vs Gemini 3.7 Flash and on par with GPT-5.6 Sol (xhigh) and Grok 4.6 (medium). Gains are driven by agentic evals (τ³-Banking, Terminal-Bench, GDPval). Discount pricing $0.75/$3.75 per 1M tokens through year-end puts it at ~$0.58/task on the intelligence–cost Pareto frontier.

Anthropic opened a Claude Team plan for scientists, starting with 10,000 seats across math, chemistry, physics, and other fields. Standard seats are free; premium seats with 5x usage limits cost $15 per month for one year, an 80% discount. Principal investigators at academic and nonprofit labs can enroll and add their groups. The expansion builds on Claude Science and the AI for Science credit program, with Anthropic saying it will grow well beyond the initial 10,000 seats.

Cline reported that Zhipu GLM-5.3 at max effort outperforms OpenAI GPT-5.6 Sol at max effort on the newly released Terminal-Bench 4.0. The post quotes Zhipu lead Zixuan Li's announcement that the benchmark itself is iterating fast enough to keep up with model releases. Cline already hosts glm-5.3 in-product: install with npm i -g cline, run /model, and select glm-5.3. ClinePass is offered as roughly 5x discounted access versus calling the direct API, continuing Cline's push to make open-weight agent models the default path rather than a hobby sidecar. For teams running terminal-native coding agents, the result matters because Terminal-Bench measures long-horizon shell work rather than single-file completion. An open-weight win against Sol on that surface changes both model routing and budget planning, especially if ClinePass pricing holds under production traffic.

OpenAI announced a 50% price reduction across its asynchronous Batch API for GPT-4o and GPT-4o-mini. The discount enables developers and enterprise engineering teams to run large-scale automated security scanning, unit test synthesis, and codebase documentation over 24-hour non-peak windows at half the standard cost.

Cognition announced stacked pricing for GPT-5.6 Sol on Devin Desktop and CLI, combining OpenAI's rate reductions with promotional discounts for an effective 76% savings through October 3.

Cloudflare introduced a 50% promotional discount for OpenAI's GPT-5.6 Sol via AI Gateway unified billing, combining edge caching, observability, and cost-effective multi-model routing.

Vercel's vibe-coding platform v0 announced a 50% discount on GPT-5.6 Sol and Fast generation modes through September 18, 2026. Developers can build full-stack React and Next.js applications and iterate on interactive UI prototypes at half the standard credit cost.

Cline reported that Zhipu AI's newly released GLM-5.3 surpassed Claude Fable and DeepSeek V4-Pro on Terminal-Bench benchmarks. Built on enhanced long-horizon agentic post-training, GLM-5.3 is accessible via ClinePass at a 5x discount ($4.99 first month), with Cline noting that open-weight agent models have reached escape velocity.