--- title: "Is Claude 3 Haiku still undominated? · Undominated.ai" canonical: https://undominated.ai/check/anthropic__claude-3-haiku/ description: "Claude 3 Haiku: whether it is still undominated. If anything in this catalogue is both better and cheaper, it is named, as of Sep 23, 2026. Unrated stays unrated." --- # Is Claude 3 Haiku still undominated? · Undominated.ai > Claude 3 Haiku: whether it is still undominated. If anything in this catalogue is both better and cheaper, it is named, as of Sep 23, 2026. Unrated stays unrated. # Is Claude 3 Haiku a good deal? Whether anything in this catalogue beats Claude 3 Haiku on both quality and price, and what you give up if it does. A computation on the current catalogue, not an opinion. 13 undominated of 136 · Sep 23, 2026 [1 · Choose & set workload](#check-input)[2 · Read verdict](#check-verdict)[3 · Inspect alternatives](#check-options) As of Sep 23, 2026, Claude 3 Haiku is dominated for Balanced on LMArena. GLM 5.3 Flash scores 277.3 higher and costs 53% less, with a covering envelope. 13 of 136 rated, priced standard models are undominated. [Inspect model evidence](/models/anthropic__claude-3-haiku/) [Compare differences & requirements](/compare/?models=anthropic%2Fclaude-3-haiku%2Cz-ai%2Fglm-5.3-flash) Current model Claude 3 Haiku Balanced 3 tokens in per 1 out Balanced Summarise Chat Code gen Agentic Monthly spend (USD) Used only to say how many months a named switching cost would take to earn back. Leave blank to skip. Switching cost (USD) GLM 5.3 Flash is both better and cheaper than Claude 3 Haiku: 277.3 points higher on LMArena and 53% less per million tokens, $0.26 cheaper at this mix. [Claude 3 Haiku](https://undominated.ai/models/anthropic__claude-3-haiku/) LMArena Elo · higher is better Scale starts at 1040 Elo ** 1194.6 Effective $/M · Balanced · lower is better ** $0.500/M [GLM 5.3 Flash](https://undominated.ai/models/z-ai__glm-5.3-flash/) LMArena Elo · higher is better Scale starts at 1040 Elo ** 1471.9 Effective $/M · Balanced · lower is better ** $0.237/M Claude 3 Haiku takes text, image, returns up to 4,096 tokens from a 200,000-token context, and is listed by 1 seller. Compared against 136 rated, priced models on this lens: 9 models dominate it and give up nothing, 5 more dominate it but give something up. Claude 3 Haiku scores 1194.6 at $0.50 per million tokens for this mix. ## Envelope-safe replacements Each row scores at least as high, costs no more, and covers this model’s context, output, modalities, tools, and reasoning. A cheaper narrower model is not listed here. | Model | LMArena | Effective $/M | You save | | --- | --- | --- | --- | | [GLM 5.3 Flash](/models/z-ai__glm-5.3-flash/) | 1471.9 +277.3 | $0.237/M | 53% | | [Gemma 4 31B](/models/google__gemma-4-31b-it/) | 1441.7 +247.1 | $0.153/M | 70% | | [Gemma 4 26B A4B](/models/google__gemma-4-26b-a4b-it/) | 1434.5 +239.9 | $0.143/M | 72% | | [GPT-5.6 Luna](/models/openai__gpt-5.6-luna/) | 1429.9 +235.3 | $0.450/M | 10% | | [MiMo-V2.5](/models/xiaomi__mimo-v2.5/) | 1427.4 +232.8 | $0.175/M | 65% | | [Qwen3.5-Flash](/models/qwen__qwen3.5-flash-02-23/) | 1397.7 +203.1 | $0.114/M | 77% | | [GPT-5.4 Nano](/models/openai__gpt-5.4-nano/) | 1372.9 +178.3 | $0.463/M | 7% | | [GPT-5 Nano](/models/openai__gpt-5-nano/) | 1319.7 +125.1 | $0.138/M | 73% | | [GPT-4.1 Nano](/models/openai__gpt-4.1-nano/) | 1284.7 +90.1 | $0.175/M | 65% | ## Higher score, lower price, named losses Not a drop-in. The loss is why these are not a recommendation. | Model | LMArena | Effective $/M | You give up | | --- | --- | --- | --- | | [Hy3](/models/tencent__hy3/) | 1440.6 | $0.144/M | image | | [DeepSeek V4 Flash 0423](/models/deepseek__deepseek-v4-flash/) | 1431.8 | $0.103/M | image | | [DeepSeek V3.2](/models/deepseek__deepseek-v3.2/) | 1424.8 | $0.302/M | context, image | | [DeepSeek V3.2 Exp](/models/deepseek__deepseek-v3.2-exp/) | 1422.6 | $0.305/M | context, image | | [Qwen3 235B A22B Instruct 2507](/models/qwen__qwen3-235b-a22b-2507/) | 1419.8 | $0.153/M | image | 136 rated, priced standard models. Chartreuse is the frontier. Cobalt is the model you named, when it is not on the staircase. Copy watch URL [Open the frontier](/frontier/) The link keeps the model and mix, using the current catalogue. Monthly spend and switching cost stay in this browser tab and are left out of the link. ## Saved decision references Save the model, workload, catalogue date and capability-preservation rule in this browser. Spend, switching cost and bill contents are not saved. No account or notifications. Constraint: replacements must preserve the model’s capabilities; any losses remain named trade-offs. Historical decisions cannot be fully replayed from saved references: past prices, scores and capability evidence are not stored. Save this reference ## Continue your investigation - [Build a shortlist](/compare/) - [Inspect alternatives](/alternatives/) - [Check billing conditions](/traps/) - [Read model evidence](/models/)