Model comparison
Claude 3 Haiku vs Kimi K2
Head-to-head evidence from 11 shared benchmark results across 5 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.
Public leaderboard positions: Claude 3 Haiku #194 (Supported); Kimi K2 #189 (Supported). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.
Evidence parity. Claude 3 Haiku and Kimi K2 share 11 comparable benchmark results. 0 of 8 categories are comparable. 1 result is unique to Claude 3 Haiku; 3 to Kimi K2.
Updated July 20, 2026- Shared results
- 11
- Claude 3 Haiku only
- 1
- Kimi K2 only
- 3
- Comparable categories
- 0 / 8
Benchmark data for Claude 3 Haiku and Kimi K2 is coming soon on BenchLM.
Confidence note. This is a partial-evidence comparison with 11 shared benchmark results across 5 evidence categories; 0 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.
Why this result
BenchLM has partial data for these models, but not enough overlapping benchmark coverage to produce a fair score-level comparison yet.
Kimi K2 is priced at $0.60 input / $2.50 output per 1M tokens, versus $0.25 input / $1.25 output per 1M tokens for Claude 3 Haiku. Claude 3 Haiku has the larger context window at 200K, compared with 128K for Kimi K2.
Category breakdown
Exact category averages are shown below. Not measured means BenchLM does not have enough sourced public coverage for that model and category.
| Category | Claude 3 Haiku | Δ | Kimi K2 |
|---|---|---|---|
| Math | Claude 3 HaikuNot measured | MarginNo overlap | Kimi K216.1 |
Operational comparison
Runtime and commercial metrics are compared only when both models have a complete sourced value.
| Metric | Claude 3 Haiku | Kimi K2 | Comparison |
|---|---|---|---|
| Input / output priceUSD per 1M tokens | Claude 3 Haiku$0.25 input / $1.25 output | Kimi K2$0.6 input / $2.5 output | Claude 3 Haiku has the lower combined listed price. |
| Generation speedtokens per second | Claude 3 Haiku138 tok/s | Kimi K243 tok/s | Claude 3 Haiku has the higher measured throughput. |
| First-answer latencyseconds to first token | Claude 3 Haiku1.16 s | Kimi K21.51 s | Claude 3 Haiku reaches the first token sooner. |
| Context windowmaximum listed tokens | Claude 3 Haiku200K | Kimi K2128K | Claude 3 Haiku lists the larger context window. |
Benchmark Deep Dive
Agentic1 benchmarks
| Benchmark | Claude 3 Haiku | Kimi K2 | Result |
|---|---|---|---|
| τ²-bench resultsSource | 21.1% | 61.1% | Kimi K2 leads |
Coding1 benchmarks
| Benchmark | Claude 3 Haiku | Kimi K2 | Result |
|---|---|---|---|
| AA-SciCodeSource | 18.6% | 34.5% | Kimi K2 leads |
Reasoning2 benchmarks
Knowledge6 benchmarks
| Benchmark | Claude 3 Haiku | Kimi K2 | Result |
|---|---|---|---|
| Artificial Analysis Intelligence IndexSource | 3.9% | 19.4% | Kimi K2 leads |
| AA-GPQA DiamondSource | 37.4% | 76.6% | Kimi K2 leads |
| AA-HLESource | 3.9% | 7.0% | Kimi K2 leads |
| AA-Omniscience IndexSource | -47.6% | -27.5% | Kimi K2 leads |
| AA-Omniscience AccuracySource | 17.2% | 26.8% | Kimi K2 leads |
| AA-Omniscience Hallucination RateSource | 78.2% | 74.2% | Kimi K2 leads |
Math2 benchmarks
Multimodal2 benchmarks
Inst. Following1 benchmarks
| Benchmark | Claude 3 Haiku | Kimi K2 | Result |
|---|---|---|---|
| AA-IFBenchSource | 36.1% | 41.5% | Kimi K2 leads |
Frequently Asked Questions (3)
Can I compare Claude 3 Haiku and Kimi K2 on BenchLM yet?
Not fully yet. BenchLM is tracking both models, but the sourced benchmark breakdown for this comparison is still coming soon.
Why does this comparison show “coming soon”?
BenchLM only shows category winners and benchmark-level calls when we have sourced results that can be compared fairly. For these models, the public benchmark coverage is not complete enough yet.
What data is available for Claude 3 Haiku and Kimi K2 today?
Claude 3 Haiku: $0.25 input / $1.25 output per 1M tokens Kimi K2: $0.60 input / $2.50 output per 1M tokens Both model pages still include creator, context window, reasoning mode, and other metadata while benchmark coverage fills in.
Related Comparisons
Choose a model with this week’s evidence
Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.
One email each week. Unsubscribe anytime.