Skip to main content

Posts

GPT-6.2 Astra Is 20% Cheaper — So Why Are Developers Calling It Unusable?

Photo by Kindel Media on Pexels OpenAI released GPT-6.2 Astra on November 17, 2026 alongside a sibling model, GPT-6.2 Sol — and with a 20% price cut from the previous Astra generation. The new API rate is $8 per million input tokens, $40 per million output tokens , down from $10/$50. Cached reads drop to $0.40/million. On paper, it's a compelling upgrade at a lower cost. Within 24 hours of launch, prominent developers were calling it "the worst model OpenAI has ever released." Both things are true simultaneously. The price cut is real. The backlash is real. Here is what actually matters depending on how you plan to use it. The Numbers at a Glance Model Input ($/1M) Output ($/1M) Cached Input Context GPT-6 Astra (prior) $10.00 $50.00 $0.50 ~1M tokens GPT-6.2 Astra (new) $8.00 $40.00 $0.40 ~1M tokens ...
Recent posts

GPT-6.2 Astra and Sol Are Cheaper — But the Reception Is Messy

Photo by Google DeepMind on Pexels OpenAI shipped GPT-6.2 Astra and Sol on November 17. The headline is a 20% price cut on Astra's API and a Sol that nearly keeps up with Astra on coding at one-fifth the cost. The sub-headline, buried in user reports: power users are calling Astra "essentially unusable" and "uneven and weird." Both things are simultaneously true, and knowing which applies to your workload matters before you migrate any production code. The Numbers That Actually Matter Model Input / Output (per 1M tokens) FrontierCode 2.0 Terminal-Bench-Science Data Analysis (Plotly run cost) GPT-6 Astra (prior) $10 / $50 — ~72% $19.97/run GPT-6.2 Astra (new) $8 / $40 — 77–78% $4.11/run (95% accuracy) GPT-6.1 Sol (prior) $2 / $10 39.0% 57.0% $2.61/run ...

Gemini 4 Hydrogen Is Out — And the Long-Context Pricing Changes the Cost Math

Google launched Gemini 4 Hydrogen on November 12, and the benchmarks are good enough to make it a real contender for your agentic pipelines — but the pricing puts it in an awkward spot against GPT-6.1 Terra. Here's how to decide whether it belongs in your stack. Photo by Tima Miroshnichenko on Pexels The Numbers That Actually Matter Let's get the prices on the table first, because that's what you'll be optimizing against at scale: Model Input (per 1M tokens) Output (per 1M tokens) Cached Input Context Gemini 4 Hydrogen $1.50 $7.50 $0.075 1M (flat) GPT-6.1 Terra $1.00 $5.00 — 272k → 2× surge Claude Haiku 4.5 $0.10 $0.50 — 100k Hydrogen costs 50% more per input token than Terra at baseline. That's a meaningful difference when you're running thousands of ag...

Perplexity's Decider v1.3 Is Priced Like a Rounding Error — Here's Why That Matters

Photo by Pavel Danilyuk on Pexels Perplexity quietly released Decider v1.3 and Decider Flash v1.3 on November 12, 2026 — and the pricing is so low it almost reads like a typo: $0.02 per million input tokens for the 27B model, $0.01/M for the 4B Flash variant, with output tokens free. If you're building any kind of agentic pipeline that needs to classify, route, rank, or extract structured data, this changes your cost math dramatically. The Numbers at a Glance Model Size Input price Output price Classification accuracy pplx-decider-v1.3 27B $0.02 / 1M Free 91.6% (63 tasks) pplx-decider-flash-v1.3 4B $0.01 / 1M Free Not separately published Claude Fable 5.2 — $8.00 / 1M $40.00 / 1M General reasoning GPT-4o (reference) — $2.50 / 1M $10.00 / 1M ...

Gemini 4 Hydrogen vs Claude Fable 5.2: A $1,300/Month Difference Worth Knowing About

Two major model releases landed within 72 hours of each other this week. Google dropped Gemini 4 Hydrogen on November 12, and Anthropic pushed Claude Fable 5.2 on November 9. Both are positioned as faster, more cost-efficient upgrades in their respective families — but they're solving different problems at very different price points. Here's how to think about which one you should actually be routing traffic to. Photo by Google DeepMind on Pexels The Numbers Side by Side Before we get into use cases, here's the raw spec comparison: Spec Gemini 4 Hydrogen Claude Fable 5.2 Release date Nov 12, 2026 Nov 9, 2026 Input price (per 1M tokens) $1.50 $8.00 Output price (per 1M tokens) $7.50 $40.00 Context window 1,048,576 tokens 1,000,000 tokens Max output Not disclosed 128,000 tok...

Claude Gets a Full Workspace — Sheets, Whiteboards, and App Connectors Explained

Photo by Matheus Bertelli on Pexels Anthropic shipped three new products this week — Claude Sheets , Claude Whiteboards , and Claude Data — along with a set of interactive connectors that let partner apps like Adobe, Atlassian, and Box run inside Claude. All of it landed ahead of Anthropic's expected IPO, and it's a clear play for the same office real estate Google Workspace and Microsoft 365 currently own. Whether it's a serious challenger or a clever demo is a different question. The Three New Products, Quickly Claude Sheets is a collaborative spreadsheet editor. It supports multiple tabs, formulas, formatting, and dropdown lists, with Claude co-editing alongside you in real time. The pitch: instead of writing a formula and hoping it's right, you describe what you want and Claude writes it — or flags when your formula logic is off. Claude Whiteboards is a shared digital board for teams. You can dump ideas, have Claude group and organize them...

Claude Fable 5.2 vs Gemini 4 Carbon: A Pricing and Benchmark Breakdown

Two new frontier models dropped within days of each other this week — Anthropic's Claude Fable 5.2 (Nov 9) and Google's Gemini 4 Carbon. Both are bidding for the top spot on the leaderboard, but they're priced very differently, and they win on different tasks. Here's the practical breakdown for developers deciding which to route their workloads to. Photo by Tara Winstead on Pexels The Pricing Reality Check Before benchmarks, costs. Both models just moved the market: Model Input (per 1M tokens) Output (per 1M tokens) Promo ends Claude Fable 5.2 $8.00 $40.00 — Gemini 4 Carbon $2.00 $10.00 Nov 30, 2026 Gemini 4 Carbon (post-promo) $4.00 $20.00 Dec 1+ Claude Fable 5.1 (previous) ~$10.00 ~$50.00 — Fable 5.2 represents a ~20% price cut from Fable 5.1....