Agentic
Directional only- GPT-5.3 Codex
- 71.4
- Kimi K2.5 (Reasoning)
- 55.0
- Weighted basis
- 2 vs 2 rows
- Reading
- Directional only
GPT-5.3 Codex vs Kimi K2.5 (Reasoning)
Updated September 3, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.
Decision reading
GPT-5.3 Codex has the higher public score estimate, 65.52 versus 60.49, but the 90% score intervals overlap. Treat that as a lead, not a settled winner.
4 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
GPT-5.3 Codex
GPT-5.3 Codex has the larger documented context window.
Confidence: documented
Chat turn cost
1K fresh input + 500 output tokens
Kimi K2.5 (Reasoning)
Kimi K2.5 (Reasoning) has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Kimi K2.5 (Reasoning)
Kimi K2.5 (Reasoning) has the lower estimated token cost for this stated workload. GPT-5.3 Codex has no published cached-input rate, so cached tokens use its listed input rate. Kimi K2.5 (Reasoning) has no published cached-input rate, so cached tokens use its listed input rate.
Confidence: rate-fallback
Show secondary and unsupported calls
Repository review cost
50K fresh input + 3K output tokens
Kimi K2.5 (Reasoning)
Kimi K2.5 (Reasoning) 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
The category averages use different weighted benchmark sets, so they are directional rather than like-for-like.
Confidence: limited
Agentic work
Tool use, computer use, and multi-step task completion
Not enough matched evidence
The category averages use different weighted benchmark sets, so they are directional rather than like-for-like.
Confidence: limited
What is actually comparable
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
4 GPT-5.3 Codex only
4 Shared
5 Kimi K2.5 (Reasoning) only
- Shared results
- 4
- GPT-5.3 Codex only
- 4
- Kimi K2.5 (Reasoning) only
- 5
- Like-for-like categories
- 0 / 8
2 categories use different evidence sets. Those rows remain visible for coverage context but do not name a winner.
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.
Coding
Directional only- GPT-5.3 Codex
- 67.2
- Kimi K2.5 (Reasoning)
- 76.8
- Weighted basis
- 3 vs 1 rows
- Reading
- Directional only
Reasoning
Not comparable- GPT-5.3 Codex
- Not measured
- Kimi K2.5 (Reasoning)
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Knowledge
Not comparable- GPT-5.3 Codex
- Not measured
- Kimi K2.5 (Reasoning)
- 87.2
- Weighted basis
- 0 vs 2 rows
- Reading
- Not comparable
Math
Not comparable- GPT-5.3 Codex
- Not measured
- Kimi K2.5 (Reasoning)
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Multilingual
Not comparable- GPT-5.3 Codex
- Not measured
- Kimi K2.5 (Reasoning)
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Multimodal
Not comparable- GPT-5.3 Codex
- Not measured
- Kimi K2.5 (Reasoning)
- 78.5
- Weighted basis
- 0 vs 1 rows
- Reading
- Not comparable
Instruction following
Not comparable- GPT-5.3 Codex
- Not measured
- Kimi K2.5 (Reasoning)
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
| Category | GPT-5.3 Codex | Kimi K2.5 (Reasoning) | Weighted basis | Reading |
|---|---|---|---|---|
| Agentic | 71.4 | 55.0 | Directional only2 vs 2 rows | Directional only |
| Coding | 67.2 | 76.8 | Directional only3 vs 1 rows | Directional only |
| Reasoning | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
| Knowledge | Not measured | 87.2 | Not comparable0 vs 2 rows | Not comparable |
| Math | 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 | 78.5 | Not comparable0 vs 1 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.
Terminal-Bench 2.0
Agentic
GPT-5.3 Codex: 77.3%Kimi K2.5 (Reasoning): 50.8%Normalized gap 26.5GPT-5.3 Codex source Kimi K2.5 (Reasoning) sourceSWE-bench Verified
Coding
GPT-5.3 Codex: 85%Kimi K2.5 (Reasoning): 76.8%Normalized gap 8.2GPT-5.3 Codex source Kimi K2.5 (Reasoning) source
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
- GPT-5.3 Codex
- $0.00875
- Fits in one request
- Kimi K2.5 (Reasoning)
- $0.0021
- Fits in one request
Kimi K2.5 (Reasoning) has the lower modeled cost
Costs use the listed standard API rates.
Repository review
50K fresh input + 3K output tokens
- GPT-5.3 Codex
- $0.1295
- Fits in one request
- Kimi K2.5 (Reasoning)
- $0.039
- Fits in one request
Kimi K2.5 (Reasoning) has the lower modeled cost
Costs use the listed standard API rates.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- GPT-5.3 Codex
- $0.525
- Fits in one request
- Cached input priced at the published list-input rate
- Kimi K2.5 (Reasoning)
- $0.162
- Fits in one request
- Cached input priced at the published list-input rate
Kimi K2.5 (Reasoning) has the lower modeled cost
GPT-5.3 Codex has no published cached-input rate, so cached tokens use its listed input rate. Kimi K2.5 (Reasoning) 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.
SpecificationGPT-5.3 CodexKimi K2.5 (Reasoning)
Context window
Maximum documented context; output-token limits may be lower.
GPT-5.3 Codex
Kimi K2.5 (Reasoning)
256K
API model ID
GPT-5.3 Codex
gpt-5.3-codex
OpenAI GPT-5.3 Codex model documentationKimi K2.5 (Reasoning)
Not sourced
Cached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
GPT-5.3 Codex
Not published
Kimi K2.5 (Reasoning)
Not published
Documented inputs
GPT-5.3 Codex
text, image
OpenAI model catalogKimi K2.5 (Reasoning)
Not sourced
Documented outputs
GPT-5.3 Codex
Kimi K2.5 (Reasoning)
Not sourced
Provider availability
GPT-5.3 Codex
Generally Available · OpenAI Responses API
OpenAI model catalogKimi K2.5 (Reasoning)
Not sourced
Reasoning profile
GPT-5.3 Codex
Reasoning
Kimi K2.5 (Reasoning)
Reasoning
Weight access
GPT-5.3 Codex
Proprietary
Kimi K2.5 (Reasoning)
Proprietary
License
GPT-5.3 Codex
Proprietary
Kimi K2.5 (Reasoning)
Proprietary
Release date
GPT-5.3 Codex
2026-02-05
Kimi K2.5 (Reasoning)
2026-02-01
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
- GPT-5.3 Codex has the higher public score estimate, 65.52 versus 60.49, but the 90% score intervals overlap.
- Workload cost
- Repository review: $0.1295 vs $0.039. Cache-heavy agent loop: $0.525 vs $0.162.
- Context tradeoff
- GPT-5.3 Codex has the larger documented window (400K).
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 evidence13 rows
Agentic
Terminal-Bench 2.0
GPT-5.3 Codex77.3%SourceKimi K2.5 (Reasoning)50.8%SourceGPT-5.3 Codex leads this result
OSWorld-Verified
GPT-5.3 Codex64.7%SourceKimi K2.5 (Reasoning)—Not directly comparable
Gert Labs
Shared sourceGPT-5.3 Codex57.47%Kimi K2.5 (Reasoning)32.58%GPT-5.3 Codex leads this result
JobBench
GPT-5.3 Codex33.7%SourceKimi K2.5 (Reasoning)—Not directly comparable
BrowseComp
GPT-5.3 Codex—Kimi K2.5 (Reasoning)60.6%SourceNot directly comparable
Coding
SWE-bench Verified
GPT-5.3 Codex85%SourceKimi K2.5 (Reasoning)76.8%SourceGPT-5.3 Codex leads this result
SWE-bench Pro
GPT-5.3 Codex56.8%SourceKimi K2.5 (Reasoning)—Not directly comparable
SWE-Rebench
GPT-5.3 Codex58.2%SourceKimi K2.5 (Reasoning)—Not directly comparable
Vibe Code Bench
Shared sourceGPT-5.3 Codex61.77%Kimi K2.5 (Reasoning)17.54%GPT-5.3 Codex leads this result
Knowledge
Math
AIME 2025
GPT-5.3 Codex—Kimi K2.5 (Reasoning)96.1%SourceNot directly comparable
Multimodal
MMMU-Pro
GPT-5.3 Codex—Kimi K2.5 (Reasoning)78.5%SourceNot directly comparable
Frequently asked questions
Which is better, GPT-5.3 Codex or Kimi K2.5 (Reasoning)?
GPT-5.3 Codex has the higher public score estimate, 65.52 versus 60.49, but the 90% score intervals overlap. The higher estimate is not a decisive winner because the uncertainty ranges overlap.
Which is better for coding, GPT-5.3 Codex or Kimi K2.5 (Reasoning)?
The current coding averages use different weighted benchmark sets, so BenchLM does not name a winner from them. Read the shared benchmark rows directly and test the models on the same task set.
Which is better for agentic tasks, GPT-5.3 Codex or Kimi K2.5 (Reasoning)?
The current agentic tasks averages use different weighted benchmark sets, so BenchLM does not name a winner from them. Read the shared benchmark rows directly and test the models on the same task set.
Which costs less, GPT-5.3 Codex or Kimi K2.5 (Reasoning)?
For the stated presets, chat costs $0.00875 on GPT-5.3 Codex and $0.0021 on Kimi K2.5 (Reasoning); repository review costs $0.1295 and $0.039; the cache-heavy agent loop costs $0.525 and $0.162. GPT-5.3 Codex has no published cached-input rate, so cached tokens use its listed input rate. Kimi K2.5 (Reasoning) has no published cached-input rate, so cached tokens use its listed input rate.
Which has the larger context window, GPT-5.3 Codex or Kimi K2.5 (Reasoning)?
GPT-5.3 Codex has the larger documented context window: 400K, compared with 256K.
Related comparisons
Last updated September 3, 2026
Watch GPT-5.3 Codex vs Kimi K2.5 (Reasoning)
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.