DeepSeek V4 Flash vs GPT-6 Luna
| Spec | DeepSeek V4 Flash | GPT-6 Luna |
|---|---|---|
| Respan Index | 45.4 (#44) | 49.2 (#36) |
| Blended price per 1M | $0.11 | $0.1 |
| Context window | 1M | 1.1M |
| Input / output per 1M | $0.086 / $0.18 | $0.05 / $0.25 |
| Max output | - | 128K |
| Input | text | text, vision |
| Released | 2026-07-31 | 2026-09-22 |
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 | DeepSeek V4 Flash | GPT-6 Luna |
|---|---|---|
| Coding | 29.8 DeepSeek V4 Flash | 29.6 GPT-6 Luna |
| Agents | 58.1 DeepSeek V4 Flash | 61.2 GPT-6 Luna |
| Reasoning | 46.4 DeepSeek V4 Flash | 57.2 GPT-6 Luna |
| Knowledge & language | 35.4 DeepSeek V4 Flash | 38.9 GPT-6 Luna |
| Human preference | 55.3 DeepSeek V4 Flash | 55.2 GPT-6 Luna |
Benchmarks both report
GPT-6 Luna leads on 9 of the 17 benchmarks both report. Higher is better. Each score links to where it was published.
| Benchmark | DeepSeek V4 Flash | GPT-6 Luna |
|---|---|---|
| Chess Puzzles (Epoch AI) | 32% | 36% |
| LMArena Elo | 1436 | 1441 |
| LMArena WebDev (LMArena) | 1582 | 1579 |
| FrontierMath Tier 4 v2 (Epoch AI) | 24.4% | 56.1% |
| FrontierMath Tiers 1-3 v2 (Epoch AI) | 57.5% | 78.9% |
| GPQA Diamond | 91% | 90.5% |
| LiveBench Agentic Coding (LiveBench) | 46.8% | 51.2% |
| LiveBench Coding (LiveBench) | 75% | 79% |
| LiveBench Data Analysis (LiveBench) | 79.3% | 73.4% |
| LiveBench Instruction Following (LiveBench) | 65.5% | 55.9% |
| LiveBench Language (LiveBench) | 79.2% | 73.8% |
| LiveBench Mathematics (LiveBench) | 86.8% | 89.1% |
| LiveBench Reasoning (LiveBench) | 86.6% | 81.8% |
| Mystery Game Puzzles (Epoch AI) | 34% | 7% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 94.4% | 98.9% |
| SimpleQA Verified | 33.6% | 41.4% |
| Terminal-Bench 4.0 (Vals AI) | 18.7% | 13.6% |
FAQ
Is DeepSeek V4 Flash better than GPT-6 Luna?
GPT-6 Luna scores 49.2 on the Respan Index (#36), ahead of DeepSeek V4 Flash at 45.4 (#44). GPT-6 Luna leads on 9 of the 17 benchmarks both report.
Which is cheaper, DeepSeek V4 Flash or GPT-6 Luna?
DeepSeek V4 Flash lists at $0.086 input and $0.18 output per 1M tokens; GPT-6 Luna at $0.05 and $0.25. GPT-6 Luna is 1.1x cheaper per token at list price.
Which has the larger context window?
DeepSeek V4 Flash accepts 1M tokens and GPT-6 Luna 1.1M tokens.