Math
Like-for-like- Claude Haiku 4.5
- 4.9
- Kimi K2
- 16.1
- Weighted basis
- 2 vs 2 rows
- Reading
- Kimi K2 leads
Claude Haiku 4.5 vs Kimi K2
Updated September 3, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.
Decision reading
Claude Haiku 4.5 has the higher public score, 57.72 versus 26.28, and the 90% score intervals do not overlap.
2 results are shared. Category rows based on different benchmark sets are marked directional and do not name a winner.
Which one for your work
Recommendations appear only when a shared evidence basis or an explicit operating constraint supports the call. Secondary and unsupported use cases stay disclosed below the initial list.
Long documents
Prompts that approach the documented context limit
Claude Haiku 4.5
Claude Haiku 4.5 has the larger documented context window.
Confidence: documented
Chat turn cost
1K fresh input + 500 output tokens
Kimi K2
Kimi K2 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
Show secondary and unsupported calls
Repository review cost
50K fresh input + 3K output tokens
Kimi K2
Kimi K2 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
Coding work
Code generation, repair, and software-engineering tasks
Not enough matched evidence
No shared weighted benchmark basis supports a winner.
Confidence: limited
Agentic work
Tool use, computer use, and multi-step task completion
Not enough matched evidence
No shared weighted benchmark basis supports a winner.
Confidence: limited
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Not enough matched evidence
The page does not recommend a cost winner because at least one model cannot fit the stated workload in one request. Claude Haiku 4.5 does not fit this workload in one request. Kimi K2 does not fit this workload in one request. Kimi K2 has no published cached-input rate, so cached tokens use its listed input rate.
Confidence: rate-fallback
What is actually comparable
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
3 Claude Haiku 4.5 only
2 Shared
- Shared results
- 2
- Claude Haiku 4.5 only
- 3
- Kimi K2 only
- 0
- Like-for-like categories
- 1 / 8
Category results, on a stated basis
Each row states whether both averages use the same weighted benchmark set. Directional and not-comparable rows remain visible, but they never receive a winner in this template.
Agentic
Not comparable- Claude Haiku 4.5
- Not measured
- Kimi K2
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Coding
Not comparable- Claude Haiku 4.5
- 73.3
- Kimi K2
- Not measured
- Weighted basis
- 1 vs 0 rows
- Reading
- Not comparable
Reasoning
Not comparable- Claude Haiku 4.5
- Not measured
- Kimi K2
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Knowledge
Not comparable- Claude Haiku 4.5
- Not measured
- Kimi K2
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Multilingual
Not comparable- Claude Haiku 4.5
- Not measured
- Kimi K2
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Multimodal
Not comparable- Claude Haiku 4.5
- Not measured
- Kimi K2
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Instruction following
Not comparable- Claude Haiku 4.5
- Not measured
- Kimi K2
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
| Category | Claude Haiku 4.5 | Kimi K2 | Weighted basis | Reading |
|---|---|---|---|---|
| Math | 4.9 | 16.1 | Like-for-like2 vs 2 rows | Kimi K2 leads |
| Agentic | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
| Coding | 73.3 | Not measured | Not comparable1 vs 0 rows | Not comparable |
| Reasoning | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
| Knowledge | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
| Multilingual | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
| Multimodal | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
| Instruction following | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
Shape of the matched evidence
Only shared public evidence is shown. Sparse evidence stays a ruled list rather than being closed into a radar shape.
Too few matched category axes support a radar. The ruled list below shows only shared benchmark results; positions use each benchmark’s normalized display scale when available.
FrontierMath v2 (Tiers 1-3)
Math
FrontierMath v2 (Tier 4)
Math
What each workload costs
Three fixed token mixes turn per-token rates into comparable decisions. Each scenario states context fit and whether cached input had to fall back to the published list-input rate.
Chat turn
1K fresh input + 500 output tokens
- Claude Haiku 4.5
- $0.0035
- Fits in one request
- Kimi K2
- $0.00185
- Fits in one request
Kimi K2 has the lower modeled cost
Costs use the listed standard API rates.
Repository review
50K fresh input + 3K output tokens
- Claude Haiku 4.5
- $0.065
- Fits in one request
- Kimi K2
- $0.0375
- Fits in one request
Kimi K2 has the lower modeled cost
Costs use the listed standard API rates.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- Claude Haiku 4.5
- $0.09
- Does not fit in one request
- Kimi K2
- $0.157
- Does not fit in one request
- Cached input priced at the published list-input rate
Claude Haiku 4.5 does not fit this workload in one request. Kimi K2 does not fit this workload in one request. Kimi K2 has no published cached-input rate, so cached tokens use its listed input rate.
Specification differences
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
SpecificationClaude Haiku 4.5Kimi K2
Context window
Maximum documented context; output-token limits may be lower.
Claude Haiku 4.5
Kimi K2
128K
API model ID
Claude Haiku 4.5
claude-haiku-4-5-20251001
Claude API pricingKimi K2
Not sourced
Cached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
Claude Haiku 4.5
$0.1 per 1M cached input tokens
Claude API pricingKimi K2
Not published
Documented inputs
Claude Haiku 4.5
Not sourced
Kimi K2
Not sourced
Documented outputs
Claude Haiku 4.5
Not sourced
Kimi K2
Not sourced
Provider availability
Claude Haiku 4.5
Not sourced
Kimi K2
Not sourced
Reasoning profile
Claude Haiku 4.5
Non-Reasoning
Kimi K2
Non-Reasoning
Weight access
Claude Haiku 4.5
Proprietary
Kimi K2
Proprietary
License
Claude Haiku 4.5
Proprietary
Kimi K2
Proprietary
Release date
Claude Haiku 4.5
2025-10-15
Kimi K2
2025-07-01
If you already use one of these models
- Deployment change
- The models list different providers, so authentication, endpoint behavior, limits, and feature support may change.
- Quality signal
- Claude Haiku 4.5 has the higher public score, 57.72 versus 26.28, and the 90% score intervals do not overlap.
- Workload cost
- Repository review: $0.065 vs $0.0375. Cache-heavy agent loop: $0.09 vs $0.157.
- Context tradeoff
- Claude Haiku 4.5 has the larger documented window (200K).
Run the same representative tasks against both endpoints before changing production traffic.
Benchmark evidence
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Browse raw public benchmark evidence5 rows
Agentic
JobBench
Claude Haiku 4.516.0%SourceKimi K2—Not directly comparable
Coding
Math
FrontierMath v2 (Tiers 1-3)
Shared sourceClaude Haiku 4.55.903%Kimi K221.404%Kimi K2 leads this result
FrontierMath v2 (Tier 4)
Shared sourceClaude Haiku 4.52.083%Kimi K20.000%Claude Haiku 4.5 leads this result
Frequently asked questions
Which is better, Claude Haiku 4.5 or Kimi K2?
Claude Haiku 4.5 has the higher public score, 57.72 versus 26.28, and the 90% score intervals do not overlap. That is the clearest overall quality signal in the current public data.
Which is better for coding, Claude Haiku 4.5 or Kimi K2?
The published evidence does not provide a shared weighted coding basis for both models, so BenchLM does not name a coding winner.
Which is better for agentic tasks, Claude Haiku 4.5 or Kimi K2?
The published evidence does not provide a shared weighted agentic tasks basis for both models, so BenchLM does not name a agentic tasks winner.
Which costs less, Claude Haiku 4.5 or Kimi K2?
For the stated presets, chat costs $0.0035 on Claude Haiku 4.5 and $0.00185 on Kimi K2; repository review costs $0.065 and $0.0375; the cache-heavy agent loop costs $0.09 and $0.157. Claude Haiku 4.5 does not fit this workload in one request. Kimi K2 does not fit this workload in one request. Kimi K2 has no published cached-input rate, so cached tokens use its listed input rate.
Which has the larger context window, Claude Haiku 4.5 or Kimi K2?
Claude Haiku 4.5 has the larger documented context window: 200K, compared with 128K.
Related comparisons
Last updated September 3, 2026
Watch Claude Haiku 4.5 vs Kimi K2
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.