Coding ranking

Best AI models for coding

Ranked language models with at least one published result on a public coding board (Arena coding, SWE-bench), ordered by their ModelCap Index position with each coding score shown beside it.

Snapshot as of 11 August 2026

Claude Fable 5 from Anthropic leads the best-for-coding list as of 11 August 2026 with Arena coding 1554 on the public coding boards, holding ModelCap Index position #1; 76 models qualify.

Models on this list

With openly downloadable weights

With a listed API price

Best AI models for coding (76)

Every row is a current, canonical language model with a public position on the ModelCap Index; the list filters by a published fact and orders by a published figure.

Best AI models for coding: current, canonical language models with a public ModelCap Index position and at least one published coding-family benchmark result, ordered by ModelCap Index rank
#	Model	Coding boards	ModelCap Index	Input $/1M	Output $/1M	Context	Weights
1	Claude Fable 5Anthropic	Arena coding 1554	#183.1	$10.00	$50.00	1M	API only
2	Claude Opus 5Anthropic	Arena coding 1530	#279.6	$5.00	$25.00	1M	API only
3	Muse Spark 1.2Meta	Arena coding 1532	#378.5	$1.25	$4.25	1M	API only
4	Kimi K3Moonshot AI	Arena coding 1541	#478.4	$3.00	$15.00	1M	Restricted license
5	Qwen3.8 MaxQwen	Arena coding 1528	#578.4	$2.00	$6.00	1M	API only
6	Gemini 3.1 Pro PreviewGoogle	Arena coding 1521	#677.7	$2.00	$12.00	1M	API only
7	GPT-5.6 SolOpenAI	Arena coding 1528	#777.3	$5.00	$30.00	1M	API only
8	Gemini 3.6 FlashGoogle	Arena coding 1523	#876.2	$1.50	$7.50	1M	API only
9	GPT-5.2 ChatOpenAI	Arena coding 1515	#974.1	$1.75	$14.00	128K	API only
10	GPT-5.5OpenAI	Arena coding 1520	#1073.9	$5.00	$30.00	1M	API only
11	Grok 4.20 Multi-AgentxAI	Arena coding 1508	#1171.8	$1.25	$2.50	2M	API only
12	Grok 4.5xAI	Arena coding 1522	#1270.7	$2.00	$6.00	500K	API only
13	GLM 5.2Z.ai	Arena coding 1506	#1370.5	$0.76	$2.42	1M	Open weights
14	Claude Sonnet 5Anthropic	Arena coding 1523	#1470.0	$2.00	$10.00	1M	API only
15	GPT-5.6 TerraOpenAI	Arena coding 1518	#1569.5	$1.00	$6.00	1M	API only
16	Qwen3.7 PlusQwen	Arena coding 1507	#1666.4	$0.32	$1.28	1M	API only
17	Gemini 3.5 Flash LiteGoogle	Arena coding 1504	#1766.0	$0.30	$2.50	1M	API only
18	Hy3Tencent	Arena coding 1501	#1864.9	$0.132	$0.528	262K	Open weights
19	GPT-5.6 LunaOpenAI	Arena coding 1498	#1963.3	$0.10	$0.60	1M	API only
20	Gemma 4 31BGoogle	Arena coding 1499	#2062.8	$0.10	$0.34	262K	Open weights
21	GPT-5.4 MiniOpenAI	Arena coding 1497	#2162.4	$0.75	$4.50	400K	API only
22	Qwen3.5 397B A17BQwen	Arena coding 1491	#2261.6	$0.50	$3.60	262K	Open weights
23	MiniMax M3MiniMax	Arena coding 1497	#2361.3	$0.30	$1.20	1M	Restricted license
24	InklingThinking Machines	Arena coding 1494	#2460.7	$0.95	$4.05	1M	Open weights
25	Gemma 4 26B A4BGoogle	Arena coding 1481	#2559.2	$0.12	$0.40	262K	Open weights
26	GLM 5V TurboZ.ai	Arena coding 1490	#2658.5	$1.20	$4.00	203K	API only
27	Kimi K2 ThinkingMoonshot AI	Arena coding 1486 · SWE-bench 63.4%	#2757.2	$0.60	$2.50	262K	Restricted license
28	o3OpenAI	Arena coding 1460	#2955.4	$2.00	$8.00	200K	API only
29	Mistral Medium 3.5Mistral AI	Arena coding 1479	#3054.4	$1.50	$7.50	262K	API only
30	Qwen3 235B A22B Instruct 2507Qwen	Arena coding 1472	#3153.6	$0.09	$0.55	262K	Open weights
31	Qwen3.5-122B-A10BQwen	Arena coding 1459	#3351.8	$0.29	$2.40	262K	Open weights
32	Qwen3 VL 235B A22B InstructQwen	Arena coding 1464	#3451.5	$0.21	$1.90	262K	Open weights
33	Mistral Large 3 2512Mistral AI	Arena coding 1468	#3551.2	$0.50	$1.50	262K	API only
34	Claude Haiku 4.5Anthropic	Arena coding 1479	#3651.0	$1.00	$5.00	200K	API only
35	GPT-5.4 NanoOpenAI	Arena coding 1460	#3946.6	$0.20	$1.25	400K	API only
36	Qwen3 Next 80B A3B InstructQwen	Arena coding 1446	#4046.6	$0.09	$1.10	262K	Open weights
37	Qwen3 VL 235B A22B ThinkingQwen	Arena coding 1455	#4245.4	$0.40	$4.00	131K	Open weights
38	Qwen3 235B A22B Thinking 2507Qwen	Arena coding 1442	#4345.3	$0.23	$2.30	262K	Open weights
39	Qwen3 Coder 480B A35BQwen	Arena coding 1457	#4643.7	$0.30	$1.00	262K	Open weights
40	o4 MiniOpenAI	Arena coding 1433	#4742.5	$1.10	$4.40	200K	API only
41	Qwen3 30B A3B Instruct 2507Qwen	Arena coding 1440	#4841.6	$0.048	$0.193	262K	Open weights
42	Qwen3 235B A22BQwen	Arena coding 1446	#5139.2	$0.455	$1.82	131K	Open weights
43	GLM 4.5 AirZ.ai	Arena coding 1426	#5437.5	$0.13	$0.85	131K	Open weights
44	GLM 4.6VZ.ai	Arena coding 1417	#5836.9	$0.30	$0.90	131K	Open weights
45	Qwen3 Next 80B A3B ThinkingQwen	Arena coding 1420	#5936.7	$0.15	$1.20	262K	Open weights
46	Trinity Large ThinkingArcee AI	Arena coding 1414	#6036.5	$0.22	$0.85	262K	Restricted license
47	GLM 4.7 FlashZ.ai	Arena coding 1424	#6236.2	$0.06	$0.40	203K	Open weights
48	Gemma 3 27BGoogle	Arena coding 1358	#6533.9	$0.08	$0.45	262K	Gated access
49	Nemotron 3 SuperNVIDIA	Arena coding 1408	#6733.0	$0.085	$0.40	1M	Restricted license
50	Mistral Small 3.2 24BMistral AI	Arena coding 1412	#6832.3	$0.094	$0.25	256K	Open weights
51	Command ACohere	Arena coding 1389	#6930.5	$2.50	$10.00	256K	Gated access
52	Qwen3 32BQwen	Arena coding 1407	#7229.2	$0.08	$0.28	131K	Open weights
53	Mercury 2Inception Labs	Arena coding 1394	#7328.4	$0.25	$0.75	128K	API only
54	Nova 2 LiteAmazon	Arena coding 1395	#8027.1	$0.30	$2.50	1M	API only
55	Gemma 3 12BGoogle	Arena coding 1316	#9325.5	$0.05	$0.15	131K	Gated access
56	Qwen3 30B A3BQwen	Arena coding 1386	#18024.3	$0.12	$0.50	131K	Open weights
57	Llama 4 MaverickMeta	Arena coding 1373	#18123.5	$0.20	$0.696	1M	Gated access
58	GPT-4 TurboOpenAI	Arena coding 1347	#18721.8	$10.00	$30.00	128K	API only
59	Llama 4 ScoutMeta	Arena coding 1362	#18821.7	$0.10	$0.30	1.3M	Gated access
60	Llama 3.3 70B InstructMeta	Arena coding 1345	#19320.3	$0.10	$0.32	131K	Gated access
61	Gemma 3n 4BGoogle	Arena coding 1308	#19420.1	$0.06	$0.12	33K	Gated access
62	Nemotron 3 Nano 30B A3BNVIDIA	Arena coding 1362	#19519.9	$0.05	$0.20	262K	Restricted license
63	Olmo 3 32B ThinkAllen Institute	Arena coding 1364	#19617.5	—	—	66K	Open weights
64	Granite 4.1 8BIBM Granite	Arena coding 1354	#19717.1	$0.05	$0.10	131K	Open weights
65	Qwen2.5 72B InstructQwen	Arena coding 1356	#19816.9	$0.36	$0.40	33K	Restricted license
66	Nova Pro 1.0Amazon	Arena coding 1343	#19914.8	$0.80	$3.20	300K	API only
67	Gemma 3 4BGoogle	Arena coding 1274	#20014.4	$0.05	$0.10	131K	Gated access
68	Qwen2.5 Coder 32B InstructQwen	Arena coding 1342 · SWE-bench 9.0%	#20511.5	$0.66	$1.00	33K	Open weights
69	Command R+ (08-2024)Cohere	Arena coding 1280	#20610.9	$2.50	$10.00	128K	API only
70	Phi 4Microsoft	Arena coding 1306	#2089.2	$0.07	$0.14	16K	Open weights
71	Command R (08-2024)Cohere	Arena coding 1281	#2117.8	$0.15	$0.60	128K	API only
72	Nova Micro 1.0Amazon	Arena coding 1288	#2137.3	$0.035	$0.14	128K	API only
73	Mixtral 8x22B InstructMistral AI	Arena coding 1277	#2155.4	$2.00	$6.00	66K	Open weights
74	Llama 3.1 8B InstructMeta	Arena coding 1260	#2193.7	$0.05	$0.08	131K	Gated access
75	Llama 3.2 3B InstructMeta	Arena coding 1176	#2260.0	$0.05	$0.33	131K	Gated access
76	Llama 3.2 1B InstructMeta	Arena coding 1149	#2270.0	$0.027	$0.201	60K	Gated access

Best AI models for coding: common questions

Which model tops the best-for-coding list right now?

Claude Fable 5 (Anthropic), as of 11 August 2026, with Arena coding 1554 on the public coding boards and ModelCap Index position #1.

How is the best-for-coding list built?

It contains current, canonical language models with a public ModelCap Index position and at least one published coding-family benchmark result, ordered by ModelCap Index rank. Nothing is estimated for the list itself: the filter is a published fact and the order is a published figure, and every row links to the model page that shows its evidence.

How many models qualify for the best-for-coding list?

76 models qualify. The population is the live ModelCap Index, so a model appears here the moment it holds a public position and meets the filter.

What is the cheapest model on the best-for-coding list?

Llama 3.1 8B Instruct lists the lowest blended API price at $0.057 per 1M tokens ($0.05 input / $0.08 output).

Which is the best open-weight model on this list?

GLM 5.2 from Z.ai is the highest-placed model with openly downloadable weights (#13 here, ModelCap Index #13).

How current is the best-for-coding list?

It is rendered from the sealed dataset published 11 August 2026 and re-renders within a minute of each refresh; positions, prices and context figures are the ones shown on the live rankings at the same instant.