Inworld optimizes Gemma 4 26B inference and cuts prices
What happened
Inworld has launched an optimized Gemma 4 26B inference service on its own platform, claiming time to first token up to 8.6x faster with a p95 of 0.23s. The service targets consumer-facing applications, with the goal of making them respond faster and cost less. Pricing per 1 million tokens is $0.07 for input, $0.34 for output and $0.035 for cached input; the cache discount applies only to context that hits the cache, not to all input.
AI-written from the coverage · updated 2 d ago
Coverage timeline
Follow the coverage to see the story from different sides.
- Inworld AIPickInworld optimizes Gemma 4 26B, bringing time to first token down to as low as 0.23s and input prices down nearly 40%
Inworld optimized Gemma 4 26B on its own platform, saying time to first token is up to 8.6x faster with a p95 of 0.23s, compared with a median of 1.97s across 15 providers serving the same model on OpenRouter (data from November 16). Input pricing is nearly 40% lower than the median provider for the same model, and cached context brings further savings. Per 1 million tokens, the price is $0.07 for input, $0.34 for output, and $0.035 for cached input.
Heat trend for this story
- Current (comparable)
- 3
- Peak (comparable)
- 10Nov 20 03:00
- 24-hour change
- -51%
The trend only compares participants observed continuously and completely, so it may cover fewer than the current heat count. Hover over or tap the chart to see each hour's heat; with a keyboard, use the left and right arrow keys.