A New Name at the Top of Text Generation
For the first time, a model called inkling-small has claimed the #1 spot in text generation, overtaking claude-fable-5 — and meanwhile, xAI is quietly expanding its image footprint with grok-imagine-image-2.0 landing in the top 10 of both Image Editing and Text-to-Image simultaneously. Three new model entries across the leaderboards make today one of the more eventful ranking days we’ve seen this month.
What Changed Today
- 🏆 New #1 in Text Generation: inkling-small overtook claude-fable-5 to claim the top spot (ELO: 1430).
- 📥 New in Text Generation Top 10: inkling-small enters the rankings for the first time (ELO: 1430).
- 📥 New in Image Editing Top 10: grok-imagine-image-2.0 (low) enters the rankings (ELO: 1439).
- 📥 New in Text-to-Image Top 10: grok-imagine-image-2.0 (low) enters the rankings (ELO: 1320).
Text Generation: inkling-small Takes the Crown
inkling-small has debuted at the top of the text generation leaderboard with an ELO of 1430, dethroning claude-fable-5. What makes this particularly notable is the name itself — “small” — suggesting this isn’t even a frontier-scale model, yet it’s outperforming the heavyweights. We don’t yet have confirmed pricing or a public API model ID for inkling-small, so developers should watch closely for official release details.
For context, the current reference leaderboard still shows claude-opus-4-6 (ELO: 1504) and gemini-3.1-pro-preview (ELO: 1500) at the top of the established rankings. An ELO of 1430 from a “small” model entering the scene is a strong signal that a new provider may be disrupting the cost-performance curve. If inkling-small delivers frontier-adjacent quality at a fraction of the price, it could become the default choice for high-volume API workloads where claude-opus-4-6 at $5/$25 per MTok is overkill.
FREE GUIDE
Stop Writing Design Specs by Hand
Get the free visual guide: how AI tools generate GAMP 5 documentation directly from your PLC and DCS exports. Used by Life Sciences engineers who are done doing it manually.
No spam. Unsubscribe anytime.
Image Editing: Grok Climbs Higher with 2.0
grok-imagine-image-2.0 (low) from xAI has entered the Image Editing top 10 with an ELO of 1439 — and that score is significant. It places it above the current listed leader, chatgpt-image-latest-high-fidelity (ELO: 1413), suggesting a potential leadership change once vote counts stabilize. xAI already had two Grok image models in the top 10 (grok-imagine-image-pro at #5 and grok-imagine-image at #6), so this 2.0 release represents a meaningful quality jump.
The “(low)” quality tier designation is particularly interesting for developers — if this is the low-fidelity variant already scoring 1439, a high-fidelity version could be a serious threat to OpenAI’s dominance in this category. Developers currently using gpt-image-1 for editing workflows should keep an eye on xAI’s API documentation for the 2.0 model ID and pricing.
Text-to-Image: xAI Double-Dips
grok-imagine-image-2.0 (low) also entered the Text-to-Image top 10 with an ELO of 1320. While this doesn’t threaten the current leader — gpt-image-1.5-high-fidelity at ELO 1247 per the reference doc — the raw score of 1320 would actually place it well above the listed #1 if we’re comparing directly, similar to the Image Editing situation. This likely reflects newer arena voting data overtaking the reference snapshot from February.
xAI now has three models in the Text-to-Image top 10 (grok-imagine-image, grok-imagine-image-pro, and now the 2.0 variant), making it the most represented provider in image generation after Google. For developers building image generation features, xAI’s Grok image models are becoming increasingly hard to ignore — especially if the “low” tier delivers this quality at a reduced price point.
Current Leaders at a Glance
| Category | #1 Model | Provider | ELO Score |
|---|---|---|---|
| Text Generation | inkling-small 🆕 | TBD | 1430 |
| Image Editing | chatgpt-image-latest-high-fidelity* | OpenAI | 1413 |
| Text-to-Image | gpt-image-1.5-high-fidelity | OpenAI | 1247 |
*Note: grok-imagine-image-2.0 (low) entered at ELO 1439 in Image Editing, which exceeds the current listed leader’s score. Leadership may shift as vote counts accumulate. Similarly, grok-imagine-image-2.0 (low) scored 1320 in Text-to-Image, above the current #1’s 1247.
So What?
Today’s changes carry three actionable signals for builders. First, if you’re running high-volume text generation workloads, put inkling-small on your radar immediately — a “small” model topping the leaderboard could mean dramatically better cost-per-quality ratios than the current frontier options like claude-opus-4-6 or gemini-3.1-pro-preview. Second, xAI is making a serious play in image generation across both editing and creation — if you’ve been defaulting to gpt-image-1 or gemini-3-pro-image-preview, it’s time to benchmark Grok’s 2.0 image models in your pipeline, especially since even the “low” tier is scoring above established leaders. Third, the broader trend here is clear: the moat around any single provider’s lead is shrinking fast. Build your abstractions accordingly — use model routers, maintain fallback chains, and don’t hardcode a single provider into your stack. The leaderboard you see today won’t be the leaderboard you see next week.
