DeepSeek V4 Flash Vision: Images Cap at 384 Tokens
DeepSeek's vision variant of V4 Flash costs the same as the text model. Images cap at 384 input tokens, and thinking mode adds 80 input tokens per call.
Discounted LLM APIs: 13 Models Below List, GPT-5.6 Sol at Half Price
Thirteen models across five series bill below list. Which cuts are the vendor's own promotion, and which are ofox's, including Sol at half OpenAI's list.
Gemini 3.7 Flash API: $0.75 Pricing, minimal Returns 400
Gemini 3.7 Flash bills $0.75 in / $3.75 out, same as 3.6 Flash. reasoning_effort minimal now returns 400, and the cheapest legal tier costs 8.4x more.
Opus 5 vs Grok 4.6 Cost: 3.6x the Bill for 1.75x the Code
One real task, measured: Opus 5 costs 3.6x Grok 4.6 and returns 1.75x as much code, in half the wall clock. Plus the usage field that hides the bill.
GLM 5.3 API: Pricing, Endpoints, and reasoning_effort
The GLM 5.3 API costs $1.4 in / $4.4 out, ships three base URLs, and defaults reasoning_effort to max.
Grok 4.6 API Pricing: The 200K Cliff and When 4.5 Wins
Grok 4.6 costs $2/$6 per 1M tokens until your prompt hits 200K, then every token doubles. Measured against Grok 4.5, including the per-request overhead.
Pi Coding Agent Custom Provider: Setup, Cost, and 3 Fixes
Pi reports $0 for every session on a custom provider, and reasoning models 500 on Anthropic routes. Both are config, not bugs. Tested on pi 0.84.1.
DeepSeek API Price Increase: Up to 12x, Peak Hours (2026)
DeepSeek's new API prices went live 2026-08-16 16:00 UTC. Cache hits jumped 12x on V4 Pro, output 4.7x on Flash. Peak hours, new rates, and the bill math.
DeepSeek Harness (dsh): Version, Updates, Stability (2026)
DeepSeek Harness (dsh) is on 0.1.1-rc.2. Development restarted 2026-08-24 and GitHub now tags releases npm has not published. Pin before you build.
Codex "Couldn't Load Its Resources": 5 Causes to Fix (2026)
Codex extension won't start? The 26.803.41515 regression broke 5 different ways. Fixed in 26.810.41047, but the error message still hides two of them.