GPT-6 Astra vs Kimi K2.6
| Spec | GPT-6 Astra | Kimi K2.6 |
|---|---|---|
| Respan Index | 72.3 (#7) | 36.3 (#77) |
| Blended price per 1M | $20 | $0.742 |
| Context window | 1.1M | 262K |
| Input / output per 1M | $10 / $50 | $0.39 / $1.8 |
| Max output | 128K | - |
| Input | text, vision | text, vision |
| Released | 2026-09-03 | 2026-04-20 |
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 | GPT-6 Astra | Kimi K2.6 |
|---|---|---|
| Coding | 58.5 GPT-6 Astra | 13.7 Kimi K2.6 |
| Agents | 78.3 GPT-6 Astra | 44.8 Kimi K2.6 |
| Reasoning | 86.0 GPT-6 Astra | 40.2 Kimi K2.6 |
| Knowledge & language | 75.6 GPT-6 Astra | 35.0 Kimi K2.6 |
| Human preference | 60.1 GPT-6 Astra | 53.9 Kimi K2.6 |
Benchmarks both report
GPT-6 Astra leads on 21 of the 22 benchmarks both report. Higher is better. Each score links to where it was published.
| Benchmark | GPT-6 Astra | Kimi K2.6 |
|---|---|---|
| LMArena WebDev (LMArena) | 1764 | 1509 |
| LMArena Elo | 1475 | 1461 |
| LMArena Vision (LMArena) | 1281 | 1265 |
| Chess Puzzles (Epoch AI) | 76% | 20% |
| EBR-bench (Epoch AI) | 76.2% | 2.4% |
| FrontierMath Tier 4 v2 (Epoch AI) | 97.6% | 25.6% |
| FrontierMath Tiers 1-3 v2 (Epoch AI) | 93.7% | 57.2% |
| Furniture Assembly (Epoch AI) | 80% | 21.7% |
| GPQA Diamond | 96% | 90.8% |
| LiveBench Agentic Coding (LiveBench) | 57.3% | 46.9% |
| LiveBench Coding (LiveBench) | 80.4% | 78.6% |
| LiveBench Data Analysis (LiveBench) | 83% | 65.1% |
| LiveBench Instruction Following (LiveBench) | 75.6% | 64.4% |
| LiveBench Language (LiveBench) | 89.4% | 75.1% |
| LiveBench Mathematics (LiveBench) | 96.8% | 84.3% |
| LiveBench Reasoning (LiveBench) | 92.7% | 79.4% |
| Mystery Game Puzzles (Epoch AI) | 84% | 18% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 100% | 96.1% |
| SAGE (Vals AI) | 46.4% | 50.2% |
| SimpleQA Verified | 75.6% | 34.9% |
| Vending-Bench 2 (Andon Labs) | 15514.7 | 6204.57 |
| WeirdML (Håvard Tveit Ihle) | 93.6% | 55.9% |
FAQ
Is GPT-6 Astra better than Kimi K2.6?
GPT-6 Astra scores 72.3 on the Respan Index (#7), ahead of Kimi K2.6 at 36.3 (#77). GPT-6 Astra leads on 21 of the 22 benchmarks both report.
Which is cheaper, GPT-6 Astra or Kimi K2.6?
GPT-6 Astra lists at $10 input and $50 output per 1M tokens; Kimi K2.6 at $0.39 and $1.8. Kimi K2.6 is 27x cheaper per token at list price.
Which has the larger context window?
GPT-6 Astra accepts 1.1M tokens and Kimi K2.6 262K tokens.
Compare other models
- vsGPT-6.2 Astra vs GPT-6 Astra
- vsKimi K2.6 vs Kimi K2.5
- vsKimi K3 vs Kimi K2.6
- vsGPT-6 Astra vs Gemini 4 Argon
- vsGPT-6 Astra vs Phi-4
- vsGPT-6 Astra vs Claude Fable 5.1
- vsGPT-6 Astra vs Kimi K3.1 Pro
- vsGPT-6 Astra vs Claude Sonnet 5.5
- vsKimi K2.6 vs Qwen3.7-Plus
- vsKimi K2.6 vs Kimi K2.7 Code
- vsKimi K2.6 vs GLM-5.1