Code review

Results are only compared within a matching group, so a model tested one way is never ranked against a model tested another way.

Review a code change and catch the real bugs it introduces without burying them in invalid or low-value comments

Ranking groups use separate scales. Compare configurations only inside the same group; EvalRank never flattens unlike systems into one leaderboard.

Cross-benchmark consensus

One merged ranking of models by where they place across this topic's benchmark boards. Derived from the boards below, not a calibrated top set; it updates as sources are re-scraped.

Rank	Model	Score	Benchmarks
1	gpt-6.1 sol macroscopebench:medium-effort	1.241	2
2	grok 4.6 macroscopebench:medium-effort	0.969	2
3	claude opus 5.5 macroscopebench:max-effort	0.952	1
4	gpt-6 sol macroscopebench:max-effort	0.907	1
5	claude opus 5.5 macroscopebench:xhigh-effort	0.864	1
6	gpt-6.1 sol macroscopebench:max-effort	0.822	1
7	gpt-6 astra macroscopebench:max-effort	0.783	1
8	gpt-5.6 sol macroscopebench:high-effort	0.745	1
9	gemini 3.8 flash macroscopebench:medium-effort	0.744	2
10	glm 5.3 macroscopebench:max-effort	0.708	1
11	gpt-6.1 sol macroscopebench:xhigh-effort	0.673	1
12	gpt-6 sol macroscopebench:xhigh-effort	0.640	1
13	gpt-6.1 sol macroscopebench:high-effort	0.608	1
14	grok 4.6 macroscopebench:xhigh-effort	0.577	1
15	grok 4.6 macroscopebench:high-effort	0.547	1
16	claude opus 5.5 macroscopebench:medium-effort	0.534	2
17	gpt-6 sol macroscopebench:high-effort	0.519	1
18	claude opus 5.5 macroscopebench:high-effort	0.464	1
19	kimi k3 macroscopebench:max-effort	0.439	1
20	gpt-6 luna macroscopebench:max-effort	0.390	1
21	gpt-6 sol macroscopebench:medium-effort	0.373	2
22	deepseek v4.1 flash macroscopebench:max-effort	0.367	1
23	deepseek v4.1 flash macroscopebench:low-effort	0.323	1
24	gpt-6.1 sol macroscopebench:low-effort	0.302	1
25	deepseek v4.1 flash macroscopebench:high-effort	0.281	1
26	claude opus 5 macroscopebench:high-effort	0.242	1
27	glm 5.3 macroscopebench:high-effort	0.224	1
28	claude opus 5 macroscopebench:medium-effort	0.200	2
29	kimi k3 macroscopebench:high-effort	0.188	1
30	gpt-6 luna macroscopebench:xhigh-effort	0.171	1
31	gpt-5.6 sol macroscopebench:low-effort	0.155	1
32	gpt-6 luna macroscopebench:high-effort	0.139	1
33	gpt-6 luna macroscopebench:medium-effort	0.108	2
34	gpt-6 sol macroscopebench:low-effort	0.093	1
35	claude opus 5.5 macroscopebench:low-effort	0.079	1
36	claude opus 5 macroscopebench:low-effort	0.065	1
37	grok 4.6 macroscopebench:low-effort	0.051	1
38	glm 5.3 macroscopebench:low-effort	0.038	1
39	kimi k3 macroscopebench:low-effort	0.025	1
40	gemini 3.8 flash macroscopebench:low-effort	0.012	1
41	gpt-6 luna macroscopebench:low-effort	0.000	1

Agent system · Agentic

Exploratory - not a ranked winner yet

rg-code-review-agent-system-agentic-agent-system-v1

Explorer evidence

MacroscopeBench, every reasoning effort (Macroscope)

Parsed yesterday

Sources differ; shown separately, never averaged. Observed 2026-10-04T10:29:46Z; fresh until 2027-01-02T10:29:46Z.

Rank	Configuration	Capability	Uncertainty	Families
1	Claude Opus 5.5 macroscopebench:max-effort`config_e6d73883b79a458bb42954a12f9a4bd2f61b19d1a69dbedddcee57a7ccfc3a20`	100%	Unknown	1
2	GPT-6 Sol macroscopebench:max-effort`config_8355540786508f62b7649c7e41ab7c7c3840eb783dc6fa18f30fa4ebf7f69482`	91%	Unknown	1
3	Claude Opus 5.5 macroscopebench:xhigh-effort`config_8b30978b3842998bf9bbb38614b78c91c8e326d7807192a9af2c2988cda4e00f`	87%	Unknown	1
4	GPT-6.1 Sol macroscopebench:max-effort`config_78df30b801e7eb04d6ef351803a5b54130f7c6a7b4d5ff9aa9756422418d076d`	86%	Unknown	1
5	GPT-6 Astra macroscopebench:max-effort`config_a75d727f10d4a0eac19d6e01bf487c6184e8a510672183ea02a01a5fc23f3c8f`	85%	Unknown	1
6	GPT-5.6 Sol macroscopebench:high-effort`config_99f8f7f0f456b8c6d1df1fa7043c58df55379cef41fa8a486e932ad43c3dfaca`	84%	Unknown	1
7	GLM 5.3 macroscopebench:max-effort`config_483032ccc1d3f89ac4535ef6e623763c1cf3e47b0f87f7636ef89ca557e32604`	83%	Unknown	1
8	GPT-6.1 Sol macroscopebench:xhigh-effort`config_82926d8185ff4aefa885e0248125762b7364ee16ea39e3de6323d76c6420dbfb`	83%	Unknown	1
9	GPT-6 Sol macroscopebench:xhigh-effort`config_b5112dbc3d69b149f10e0a3d43ec90644a4a03cb6431d7441b6cd137f4d727e3`	83%	Unknown	1
10	GPT-6.1 Sol macroscopebench:high-effort`config_5a24f76c989220b98e8d8a772ebe10c60ec362db74a637fec8cbe6db7ddc6300`	82%	Unknown	1
11	Grok 4.6 macroscopebench:xhigh-effort`config_40f36156eaa01fa3d0fa03fe65b5d085cd5e91e64eeff83efcb02980d3ba0628`	80%	Unknown	1
12	Grok 4.6 macroscopebench:high-effort`config_9c445642dfb058b00be7b564987b4f27af2bff78275a8d43bda504fd245abf43`	78%	Unknown	1
13	GPT-6 Sol macroscopebench:high-effort`config_e25f44e408a733401428874642e5edfe2bf935f05a41052f7ab74ee0fbb99318`	76%	Unknown	1
14	GPT-6.1 Sol macroscopebench:medium-effort`config_9dafacf26cdd886ba1c1f49d87065b7d3f5cd54feb06f8ad8d6638188d37d663`	73%	Unknown	1
15	Claude Opus 5.5 macroscopebench:high-effort`config_733dd5cc9437841e5449fa06b422c17e83334f55168c2f5d5b9325a535289ffc`	72%	Unknown	1
16	Kimi K3 macroscopebench:max-effort`config_fc2b52a594c8a99a3b7c2b59e1391c3c83620d834ea81943588d979eda112be6`	72%	Unknown	1
17	Grok 4.6 macroscopebench:medium-effort`config_10f0ba0fa4acbaba4d1227648e8a00f2cb7bc77f44f64a7b4ccf1880e161daf2`	72%	Unknown	1
18	GPT-6 Luna macroscopebench:max-effort`config_3c0dd5fe33609fcaddb22f06a3e078e4f268b2003568cc116ae7eb7384151607`	71%	Unknown	1
19	DeepSeek V4.1 Flash macroscopebench:max-effort`config_1b29e838bc883e7f200532f0c15b9414fcb38a2e33e878b81ffa9203b1f651e2`	70%	Unknown	1
20	Gemini 3.8 Flash macroscopebench:medium-effort`config_9da81bf684efbe1ad1948d51b5b78cb6cd93e9dad9c50d8fc7887883e3a32b50`	66%	Unknown	1
21	DeepSeek V4.1 Flash macroscopebench:low-effort`config_ce17896acd252f51af585df90ee2ec0e3d45888ceb07c6a1dd33a711252d6ed4`	66%	Unknown	1
22	GPT-6.1 Sol macroscopebench:low-effort`config_7c8deb8f693e6c268a2313d787bf120b0ea2e5f62cd4f1417060353ee57317f8`	66%	Unknown	1
23	DeepSeek V4.1 Flash macroscopebench:high-effort`config_2b2e93dfccd56f89985469205995a700afa68ca234cb33876dd8bebc4d1b4b49`	66%	Unknown	1
24	Claude Opus 5.5 macroscopebench:medium-effort`config_81b476353cd67c47751bf44dac77ad0f9eee84f3498f3c8b89f694420f430ff3`	65%	Unknown	1
25	Claude Opus 5 macroscopebench:high-effort`config_88f804bb21b203128d39a5f2c65aac58f46eaf269947654a8c16b50f2d1a8219`	64%	Unknown	1
26	GLM 5.3 macroscopebench:high-effort`config_f8c73c8a27ed8ec1ddba264606a44cea6f03daac9b87b51d8be381f565cbeb04`	63%	Unknown	1
27	GPT-6 Sol macroscopebench:medium-effort`config_60c27177bb62cbb0bc1a1315c3195c7739acd13b0107510df793d173238ae34f`	63%	Unknown	1
28	Kimi K3 macroscopebench:high-effort`config_9197026ae1add69be5447726f323c3b6df043eaf50ff17785efd62297e0def15`	61%	Unknown	1
29	GPT-6 Luna macroscopebench:xhigh-effort`config_88e3c4fc4945be6778e224c4012ab03451b10e72c22e0ee09ea2182d94eb2703`	60%	Unknown	1
30	GPT-5.6 Sol macroscopebench:low-effort`config_accd884ef51fabc161a52d15d9b4db38cca243783d42ed7108a2f8b5d73a3402`	55%	Unknown	1
31	GPT-6 Luna macroscopebench:high-effort`config_82058abb9bbc34aa1c34ae904cca0e264290b8f3b3f4e00bd6a94f4416e17652`	53%	Unknown	1
32	Claude Opus 5 macroscopebench:medium-effort`config_88aafab5a5dd6c8836dce41a23b99df08d5a0740972946670a44b907435db23b`	53%	Unknown	1
33	GPT-6 Luna macroscopebench:medium-effort`config_29792f84f3e1cb2bc35c50d8e07675f7a7d08a8e1772251c5544842835e305ba`	45%	Unknown	1
34	GPT-6 Sol macroscopebench:low-effort`config_ef2e7cca45952dfa1c5f5bb88e5a50eb9988bf57d5be8e41595685154337e61f`	44%	Unknown	1
35	Claude Opus 5.5 macroscopebench:low-effort`config_0c2ade372c70318cb921a18ade2cfc1572dc5691bf9b7db84d5b9bb1f4c19160`	43%	Unknown	1
36	Claude Opus 5 macroscopebench:low-effort`config_db91574819230d1677188be05a134058093d98912c162f534aff9147cb2b76e9`	38%	Unknown	1
37	Grok 4.6 macroscopebench:low-effort`config_3368cc140fbfd7ce4dadd75a37bb9e79ffa5019153bd0554318a22c185f860cb`	32%	Unknown	1
38	GLM 5.3 macroscopebench:low-effort`config_888ee1fd139c4c552ceab0300b8ab5cd4d374e6a23f8b8f36006393e328f593b`	25%	Unknown	1
39	Kimi K3 macroscopebench:low-effort`config_7aab349be7a0f827ef808b8c22a1e74af5ea0f2d625d7b523b9fb39ed3d32db7`	22%	Unknown	1
40	Gemini 3.8 Flash macroscopebench:low-effort`config_02e963f4bb6cd6a43724fd3a7e10a67dc2f157f877b244978c1d858c2e24d18d`	17%	Unknown	1
41	GPT-6 Luna macroscopebench:low-effort`config_246e3f618910cca9f31cb112d2ec39ccdad228a931a1c22d0b5fd7430147faaf`	0%	Unknown	1

Compare configurations in this group

macroscopebench benchmark evidence

Source: Macroscope

Explorer evidence

MacroscopeBench, medium effort (Macroscope)

Parsed yesterday

Sources differ; shown separately, never averaged. Observed 2026-10-04T10:29:48Z; fresh until 2027-01-02T10:29:48Z.

Rank	Configuration	Capability	Uncertainty	Families
1	GPT-6.1 Sol macroscopebench:medium-effort`config_9dafacf26cdd886ba1c1f49d87065b7d3f5cd54feb06f8ad8d6638188d37d663`	100%	Unknown	1
2	Grok 4.6 macroscopebench:medium-effort`config_10f0ba0fa4acbaba4d1227648e8a00f2cb7bc77f44f64a7b4ccf1880e161daf2`	94%	Unknown	1
3	Gemini 3.8 Flash macroscopebench:medium-effort`config_9da81bf684efbe1ad1948d51b5b78cb6cd93e9dad9c50d8fc7887883e3a32b50`	75%	Unknown	1
4	Claude Opus 5.5 macroscopebench:medium-effort`config_81b476353cd67c47751bf44dac77ad0f9eee84f3498f3c8b89f694420f430ff3`	69%	Unknown	1
5	GPT-6 Sol macroscopebench:medium-effort`config_60c27177bb62cbb0bc1a1315c3195c7739acd13b0107510df793d173238ae34f`	61%	Unknown	1
6	Claude Opus 5 macroscopebench:medium-effort`config_88aafab5a5dd6c8836dce41a23b99df08d5a0740972946670a44b907435db23b`	28%	Unknown	1
7	GPT-6 Luna macroscopebench:medium-effort`config_29792f84f3e1cb2bc35c50d8e07675f7a7d08a8e1772251c5544842835e305ba`	0%	Unknown	1

Compare configurations in this group

macroscopebench-medium benchmark evidence

Source: Macroscope

Not enough independent evidence for a top set

Calibration has not been validated
Independent benchmark-family coverage is below the required threshold

Back to all kinds of work