Compare / head-to-head
Claude Haiku 4.5vsKimi K2.6
Kimi K2.6 leads 3 of 3 shared benchmarks. Kimi K2.6 has the larger context window (256K tokens).
Benchmarks from cited public sources; pricing from official pages; status from official provider feeds.
Shared benchmarks
0 – 3
Kimi K2.6 leads
Cheaper per token
—
list price, input + output
Larger context
Kimi K2.6
256K tokens
Provider uptime (30d)
100% · 100%
Anthropic · Moonshot AI
Claude Haiku 4.5
Anthropic · released 2025-10-15
textvision
Kimi K2.6
Moonshot AI · released 2026-04-20
textvision
- Context
- 256K
- Max out
- —
- Input /1M
- —
- Output /1M
- —
- Cached /1M
- —
- Scores
- 14 · 8 core
Quality
Benchmark matrix
3 shared · 1 only Claude Haiku 4.5 · 11 only Kimi K2.6| Benchmark | Claude Haiku 4.5 | Kimi K2.6 | Δ | Edge |
|---|---|---|---|---|
| Reported by both · 3 | ||||
| GPQA Diamond | 71.2% ↗epoch | 90.8% ↗epoch | -19.6 pt | Kimi K2.6 |
| OTIS Mock AIME 2024-2025 (Epoch AI run) | 66.7% ↗epoch | 96.1% ↗epoch | -29.4 pt | Kimi K2.6 |
| SimpleQA Verified | 13.2% ↗epoch | 34.9% ↗epoch | -21.7 pt | Kimi K2.6 |
| Only Claude Haiku 4.5 reports · 1 | ||||
| SWE-bench Verified | 73.3% ↗ | not reported | — | — |
| Only Kimi K2.6 reports · 11 | ||||
| AIME 2026 | not reported | 95.8% ↗matharena ⚠ | — | — |
| LMArena Elo | not reported | 1454.9 ↗ | — | — |
| FrontierMath Tier 4 v2 (Epoch AI run) | not reported | 25.6% ↗epoch | — | — |
| FrontierMath Tiers 1-3 v2 (Epoch AI run) | not reported | 57.2% ↗epoch | — | — |
| Kimi Claw 24/7 Bench | not reported | 42.9% ↗ | — | — |
| Kimi Code Bench v2 | not reported | 50.9% ↗ | — | — |
| MCP-Atlas | not reported | 69.4% ↗ | — | — |
| MCPMark-Verified | not reported | 72.8% ↗ | — | — |
| MLS-Bench-Lite | not reported | 26.7% ↗ | — | — |
| ProgramBench | not reported | 48.3% ↗ | — | — |
| SWE-bench Verified (Epoch AI run) | not reported | 76.7% ↗epoch | — | — |
Scores tagged "epoch" or "matharena" are independent runs, used only where the lab has not published its own; ⚠ marks rows MathArena flags as released after the competition. Higher is better on every row. Δ is Claude Haiku 4.5 minus Kimi K2.6 in the benchmark's own unit. "Not reported" means the lab has not published that figure; it is not a zero. ↗ opens the source.
Specs & pricing
Side by side
Official pricing pages and model cards| Spec | Claude Haiku 4.5 | Kimi K2.6 | Edge |
|---|---|---|---|
| Context window | 200K tokens | 256K tokens | Kimi K2.6 |
| Max output | 64K tokens | — | — |
| Input price / 1M | $1 ↗ | — | — |
| Output price / 1M | $5 ↗ | — | — |
| Cached input / 1M | $0.1 ↗ | — | — |
| Input + output / 1M Lower is cheaper. List prices; batch, tool and regional fees excluded. | $6 | — | — |
| Modalities | text · vision | text · vision | Tie |
| Released | 2025-10-15 | 2026-04-20 | — |
| Cited benchmark scores | 4 | 14 | — |
Reliability
Provider status
Last 60 days · refreshed daily on this page · live on /statusMore matchups
Claude Haiku 4.5 vs …
Models sharing the most benchmarksMore matchups
Kimi K2.6 vs …
Models sharing the most benchmarksBuilt by Respan
Which one wins on your data?
Public benchmarks are a starting point. Run Claude Haiku 4.5 and Kimi K2.6 on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.