xai/grok-4.3| Provider | Input, $ per 1M | Output, $ per 1M | In our data since |
|---|---|---|---|
| daoxe | $1.25 | $2.5 | 31 Jul 2026 |
| frogbot | $1.25 | $2.5 | 31 Jul 2026 |
| auriko | $1.25 | $2.5 | 31 Jul 2026 |
| ofox | $1.25 | $2.5 | 31 Jul 2026 |
| bedrock_mantle | $1.25 | $2.5 | 31 Jul 2026 |
| AIHubMix | $1.25 | $2.5 | 31 Jul 2026 |
| Abacus.AI | $1.25 | $2.5 | 31 Jul 2026 |
| Amazon Bedrock | $1.25 | $2.5 | 31 Jul 2026 |
| CrossModel | $1.25 | $2.5 | 31 Jul 2026 |
| FastRouter | $1.25 | $2.5 | 31 Jul 2026 |
| Kilo Code | $1.25 | $2.5 | 31 Jul 2026 |
| LLM Gateway | $1.25 | $2.5 | 31 Jul 2026 |
| Merge Gateway | $1.25 | $2.5 | 31 Jul 2026 |
| NanoGPT | $1.25 | $2.5 | 31 Jul 2026 |
| OpenRouter | $1.25 | $2.5 | 31 Jul 2026 |
| OrcaRouter | $1.25 | $2.5 | 31 Jul 2026 |
| Vercel | $1.25 | $2.5 | 31 Jul 2026 |
| ZenMux | $1.25 | $2.5 | 31 Jul 2026 |
| xAI | $1.25 | $2.5 | 31 Jul 2026 |
| Venice AI | $1.42 | $2.83 | 31 Jul 2026 |
Only the base tier and only per-token prices are shown. Batch, discounted and cached rates, as well as prices per image or per second of video, do not go into this table: they cannot stand in the same column as a price per million tokens.
The date in the last column is the day this price first entered our collection. The price may well be older: before that day we simply were not recording it. It has not changed since — otherwise a new row with a new date would stand in its place.
| Task set | Result | Run conditions | Measured by |
|---|---|---|---|
| Arena Score, programming | 1,415.7 1,409–1,422 | — | — |
| Arena Score, multi-turn dialogue | 1,408.48 1,401–1,416 | — | — |
| Arena Score in French | 1,407.87 1,391–1,424 | — | — |
| Arena Score in Russian | 1,406.98 1,398–1,416 | — | — |
| Arena Score in English | 1,405.67 1,400–1,411 | — | — |
| Arena Score in Spanish | 1,401.89 1,384–1,419 | — | — |
| Arena Score, hard prompts | 1,400.08 1,395–1,405 | — | — |
| Arena Score, overall | 1,399.77 1,396–1,404 | — | — |
| Arena Score, long queries | 1,397.52 1,392–1,403 | — | — |
| Arena Score, mathematics | 1,390.9 1,378–1,404 | — | — |
| Arena Score, creative writing | 1,388.05 1,380–1,396 | — | — |
| Arena Score, expert questions | 1,387.07 1,377–1,397 | — | — |
| Arena Score, instruction following | 1,367.98 1,362–1,374 | — | — |
| Arena Score, web development | 1,356.93 1,350–1,364 | — | — |
| Arena Score, understanding diagrams | 1,244.36 1,234–1,255 | — | — |
| Arena Score, text recognition in images | 1,237.65 1,230–1,245 | — | — |
| Arena Score, working with images | 1,232.65 1,225–1,240 | — | — |
| Mock AIME 2024–2025 — olympiad problems | 93.33 % | effort: high · with a tuned harness | Epoch evaluations |
| GPQA Diamond — graduate-level questions | 85.1 % | effort: high · with a tuned harness | Epoch evaluations |
| WeirdML — unusual machine learning tasks | 49.89 % | — | https://htihle.github.io/weirdml.html |
| FrontierMath, levels 1–3 | 42.81 % | effort: high | Epoch evaluations |
| SimpleQA Verified — factual accuracy | 38 % | effort: high | Epoch evaluations |
| Chess puzzles | 21.09 % | effort: high | Epoch evaluations |
| FrontierMath, level 4 — research-grade problems | 14.63 % | effort: high | Epoch evaluations |
| CritPt — physics problems | 8 % | effort: high | — |
| Arena Score, agent tasks | -0.14 0–0 | — | — |