Gemini 4 Neon vs Qwen3.8-Max
| Spec | Gemini 4 Neon | Qwen3.8-Max |
|---|---|---|
| Respan Index | 78.2 (#2) | 52.8 (#35) |
| Blended price per 1M | $6 | $2.48 |
| Context window | 1M | 1M |
| Input / output per 1M | $3 / $15 | $1.65 / $4.95 |
| Max output | 66K | 131K |
| Input | text, image, audio, video | text, vision |
| Released | 2026-11-18 | 2026-08-02 |
Respan Index by category
Each category is scored 0 to 100, and how far ahead the winner is. Leads under a point are close to even.
| Category | Gemini 4 Neon | Qwen3.8-Max |
|---|---|---|
| Coding | 74.1 Gemini 4 Neon | 39.3 Qwen3.8-Max |
| Agents | 84.5 Gemini 4 Neon | 68.6 Qwen3.8-Max |
| Reasoning | 87.8 Gemini 4 Neon | 49.4 Qwen3.8-Max |
| Knowledge & language | 75.3 Gemini 4 Neon | 48.1 Qwen3.8-Max |
| Human preference | 60.4 Gemini 4 Neon | 58.0 Qwen3.8-Max |
Benchmarks both report
Gemini 4 Neon leads on 12 of the 13 benchmarks both report. Higher is better. Each score links to where it was published.
| Benchmark | Gemini 4 Neon | Qwen3.8-Max |
|---|---|---|
| FrontierSWE V2 (Proximal Labs) | 62.4% | 17.8% |
| APEX-Agents (Mercor) | 82.9% | 63.3% |
| Chess Puzzles (Epoch AI) | 83% | 46% |
| FrontierMath Tier 4 v2 (Epoch AI) | 97.6% | 46.3% |
| FrontierMath Tiers 1-3 v2 (Epoch AI) | 91.2% | 74.7% |
| Furniture Assembly (Epoch AI) | 89.2% | 20% |
| GPQA Diamond | 96.3% | 92.6% |
| LMArena WebDev (LMArena) | 1765 | 1670 |
| Mystery Game Puzzles (Epoch AI) | 85% | 38% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 100% | 100% |
| SimpleQA Verified | 75.2% | 47.3% |
| LMArena Elo | 1513 | 1484 |
| LMArena Agent (LMArena) | 0.1841 | -0.0091 |
FAQ
Is Gemini 4 Neon better than Qwen3.8-Max?
Gemini 4 Neon scores 78.2 on the Respan Index (#2), ahead of Qwen3.8-Max at 52.8 (#35). Gemini 4 Neon leads on 12 of the 13 benchmarks both report.
Which is cheaper, Gemini 4 Neon or Qwen3.8-Max?
Gemini 4 Neon lists at $3 input and $15 output per 1M tokens; Qwen3.8-Max at $1.65 and $4.95. Qwen3.8-Max is 2.4x cheaper per token at list price.
Which has the larger context window?
Both accept 1M tokens of context.