MCP tool orchestration
Results are only compared within a matching group, so a model tested one way is never ranked against a model tested another way.
Coordinate MCP tools to complete a task
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 | muse spark 1.1 mcp-atlas:muse spark 1.1 | 0.943 | 1 |
| 2 | qwen 3.8 max qwen3.8-max_sierra | 0.931 | 1 |
| 3 | glm 5.3 flash toolathlon-verified:glm 5.3 flash | 0.923 | 1 |
| 4 | claude-fable 5.1 mcp-atlas:claude-fable 5.1 | 0.889 | 1 |
| 5 | claude opus 5 claude-opus-5_sierra | 0.867 | 1 |
| 6 | kimi k3 toolathlon-verified:kimi k3 | 0.852 | 1 |
| 7 | claude-opus-5 mcp-atlas:claude-opus-5 | 0.838 | 1 |
| 8 | grok 4.5 grok-4.5_sierra | 0.806 | 1 |
| 9 | qwen3.8.2.4t-a95b mcp-atlas:qwen3.8.2.4t-a95b | 0.789 | 1 |
| 10 | claude opus 4.8 toolathlon-verified:claude opus 4.8 | 0.786 | 1 |
| 11 | gpt-5.6-sol gpt-5.6-sol_sierra | 0.750 | 1 |
| 12 | glm 5.3 mcp-atlas:glm 5.3 | 0.744 | 1 |
| 13 | muse spark 1.2 toolathlon-verified:muse spark 1.2 | 0.724 | 1 |
| 14 | gemini 3.5 flash mcp-atlas:gemini 3.5 flash | 0.700 | 1 |
| 15 | gpt-5.5 gpt-5.5_sierra | 0.697 | 1 |
| 16 | muse spark 1.1 toolathlon-verified:muse spark 1.1 | 0.667 | 1 |
| 17 | claude fable 5 mcp-atlas:claude fable 5 | 0.659 | 1 |
| 18 | muse spark 1.1 muse-spark-1.1_sierra | 0.647 | 1 |
| 19 | kimi-k3 mcp-atlas:kimi-k3 | 0.619 | 1 |
| 20 | deepseek v4 pro 0813 toolathlon-verified:deepseek v4 pro 0813 | 0.613 | 1 |
| 21 | claude opus 4.7 claude-opus-4.7_sierra | 0.600 | 1 |
| 22 | muse spark mcp-atlas:muse spark | 0.581 | 1 |
| 23 | gpt-5.5 toolathlon-verified:gpt-5.5 | 0.563 | 1 |
| 24 | claude opus 4.8 claude-opus-4.8_sierra | 0.556 | 1 |
| 25 | claude-opus-4.8 mcp-atlas:claude-opus-4.8 | 0.545 | 1 |
| 26 | claude sonnet 5 toolathlon-verified:claude sonnet 5 | 0.515 | 1 |
| 27 | claude fable 5 claude-fable-5_sierra | 0.514 | 1 |
| 28 | gpt-5.6-sol mcp-atlas:gpt-5.6-sol | 0.511 | 1 |
| 29 | inkling-small mcp-atlas:inkling-small | 0.478 | 1 |
| 30 | gpt-5.4 gpt-5.4_sierra | 0.474 | 1 |
| 31 | deepseek v4 flash 0731 toolathlon-verified:deepseek v4 flash 0731 | 0.471 | 1 |
| 32 | claude-opus-4.7 mcp-atlas:claude-opus-4.7 | 0.447 | 1 |
| 33 | kimi k3 kimi-k3_sierra | 0.436 | 1 |
| 34 | gemini 3.5 flash toolathlon-verified:gemini 3.5 flash | 0.429 | 1 |
| 35 | gemini-3.1-pro-preview mcp-atlas:gemini-3.1-pro-preview | 0.417 | 1 |
| 36 | glm-5.2 glm-5.2_sierra | 0.400 | 1 |
| 37 | gemini 3.1 pro toolathlon-verified:gemini 3.1 pro | 0.389 | 1 |
| 38 | glm-5p2 mcp-atlas:glm-5p2 | 0.388 | 1 |
| 39 | gpt-5.2 gpt-5.2_sierra | 0.366 | 1 |
| 40 | claude-opus-4.6 mcp-atlas:claude-opus-4.6 | 0.360 | 1 |
| 41 | glm 5.2 toolathlon-verified:glm 5.2 | 0.351 | 1 |
| 42 | distyl buttonagent distyl-buttonagent_distyl | 0.333 | 1 |
| 43 | inkling mcp-atlas:inkling | 0.333 | 1 |
| 44 | kimi k2.7 code toolathlon-verified:kimi k2.7 code | 0.316 | 1 |
| 45 | glm-5p1 mcp-atlas:glm-5p1 | 0.308 | 1 |
| 46 | claude opus 4.6 claude-opus-4.6_sierra | 0.302 | 1 |
| 47 | gpt-5.5 mcp-atlas:gpt-5.5 | 0.283 | 1 |
| 48 | kimi k2.6 toolathlon-verified:kimi k2.6 | 0.282 | 1 |
| 49 | gemini 3 flash gemini-3-flash_sierra | 0.273 | 1 |
| 50 | gpt-5.4 mcp-atlas:gpt-5.4 | 0.259 | 1 |
| 51 | gemini 3.5 flash-lite toolathlon-verified:gemini 3.5 flash-lite | 0.250 | 1 |
| 52 | gemini 3.1 pro preview gemini-3.1-pro-preview_sierra | 0.244 | 1 |
| 53 | gemini-3-pro-preview mcp-atlas:gemini-3-pro-preview | 0.236 | 1 |
| 54 | hy3 toolathlon-verified:hy3 | 0.220 | 1 |
| 55 | claude sonnet 4.5 claude-sonnet-4.5_sierra | 0.217 | 1 |
| 56 | claude-opus-4.5 mcp-atlas:claude-opus-4.5 | 0.214 | 1 |
| 57 | claude-sonnet-4.6 mcp-atlas:claude-sonnet-4.6 | 0.193 | 1 |
| 58 | inkling inkling_sierra | 0.191 | 1 |
| 59 | deepseek v4 pro toolathlon-verified:deepseek v4 pro | 0.190 | 1 |
| 60 | gpt-5.2 mcp-atlas:gpt-5.2 | 0.172 | 1 |
| 61 | claude opus 4.5 claude-opus-4.5_sierra | 0.167 | 1 |
| 62 | inkling-small toolathlon-verified:inkling-small | 0.163 | 1 |
| 63 | kimi-k2p5 mcp-atlas:kimi-k2p5 | 0.153 | 1 |
| 64 | grok 4.2 grok-4.2_sierra | 0.143 | 1 |
| 65 | deepseek v4 flash toolathlon-verified:deepseek v4 flash | 0.136 | 1 |
| 66 | nemotron 3 ultra mcp-atlas:nemotron 3 ultra | 0.133 | 1 |
| 67 | gemini 3 pro gemini-3-pro_sierra | 0.120 | 1 |
| 68 | gemini-3-flash-preview mcp-atlas:gemini-3-flash-preview | 0.115 | 1 |
| 69 | mimo v2.5 toolathlon-verified:mimo v2.5 | 0.111 | 1 |
| 70 | grok 4 fast grok-4-fast_sierra | 0.098 | 1 |
| 71 | claude-sonnet-4.5 mcp-atlas:claude-sonnet-4.5 | 0.097 | 1 |
| 72 | minimax m2.7 toolathlon-verified:minimax m2.7 | 0.087 | 1 |
| 73 | glm-4p7 mcp-atlas:glm-4p7 | 0.079 | 1 |
| 74 | gemini 2.5 pro gemini-2.5-pro_sierra | 0.077 | 1 |
| 75 | inkling toolathlon-verified:inkling | 0.064 | 1 |
| 76 | gemini-3.1-flash-lite mcp-atlas:gemini-3.1-flash-lite | 0.063 | 1 |
| 77 | grok 4.1 fast grok-4.1-fast_sierra | 0.057 | 1 |
| 78 | gpt-5.4-mini mcp-atlas:gpt-5.4-mini | 0.046 | 1 |
| 79 | qwen3.5 397b-a17b toolathlon-verified:qwen3.5 397b-a17b | 0.042 | 1 |
| 80 | gpt-5.2 gpt-5.2-none_sierra | 0.037 | 1 |
| 81 | gpt-5.1 mcp-atlas:gpt-5.1 | 0.030 | 1 |
| 82 | nemotron 3 ultra toolathlon-verified:nemotron 3 ultra | 0.020 | 1 |
| 83 | glm-5 glm-5-think_sierra | 0.018 | 1 |
| 84 | o3-pro mcp-atlas:o3-pro | 0.015 | 1 |
| 85 | claude-haiku-4.5 mcp-atlas:claude-haiku-4.5 | 0.000 | 1 |
| 86 | kimi k2.5 toolathlon-verified:kimi k2.5 | 0.000 | 1 |
| 87 | qwen3.5.397b-a17b qwen3.5.397b-a17b-think_sierra | 0.000 | 1 |
Agent system ยท Agentic
Exploratory - not a ranked winner yetrg-mcp-tool-orchestration-agent-system-agentic-agent-system-v1Explorer evidence
MCP-Atlas
Sources differ; shown separately, never averaged. Observed ; fresh until .
| Rank | Configuration | Capability | Uncertainty | Families |
|---|---|---|---|---|
| 1 | Muse Spark 1.1 mcp-atlas:Muse Spark 1.1config_865e6ef688acad27a266666361d822a7e12a1942834149471cf90eb044a5cc19 | 100% | Unknown | 1 |
| 2 | claude-Fable 5.1 mcp-atlas:claude-Fable 5.1config_f328e3998ed1b45103e1649d2a56eb49ab52c4778c1c7fbb601e4a09c6aeaa84 | 98% | Unknown | 1 |
| 3 | claude-opus-5 mcp-atlas:claude-opus-5 | xhighconfig_3b3da0e8d86c658bad97c185bb9a483596e0071070d7cf422d13e9cc9fe0b05a | 95% | Unknown | 1 |
| 4 | Qwen3.8-2.4T-A95B mcp-atlas:Qwen3.8-2.4T-A95B | xhighconfig_1390b81eca2217ffd224a50bf934dcc42987b13c9d626c93327d94eb0f77af78 | 92% | Unknown | 1 |
| 5 | GLM 5.3 mcp-atlas:GLM 5.3config_4230f907d1dac8530ad295039e60329f300e8b5c03fbdce4de791a419dedd214 | 92% | Unknown | 1 |
| 6 | Gemini 3.5 Flash mcp-atlas:Gemini 3.5 Flash | highconfig_b5619ab8c5a5e56ed4fb07192538b4c60398967a98bd1e84aa003fd40b5e4701 | 91% | Unknown | 1 |
| 7 | Claude Fable 5 mcp-atlas:Claude Fable 5config_0af42acc3be8b535fe45e8528a110115e3dd4285b0a8d6240ec15d79e7e0a663 | 90% | Unknown | 1 |
| 8 | kimi-k3 mcp-atlas:kimi-k3 | maxconfig_6f65c5cb596ddad3df6a940b2b8290cd820b7461521faa187b025da51bb12bbc | 88% | Unknown | 1 |
| 9 | Muse Spark mcp-atlas:Muse Sparkconfig_47028847b57705edc5e8485a9c561b1f1088e683eee397c8f02e5aca140727c7 | 88% | Unknown | 1 |
| 10 | claude-opus-4-8 mcp-atlas:claude-opus-4-8 | maxconfig_56ece83507303ffc5a54b226e22f8386a83c748a7904b50812c53e8ad254d616 | 88% | Unknown | 1 |
| 11 | gpt-5.6-sol mcp-atlas:gpt-5.6-solconfig_14e638ae9ebf6ea3d9836679b084df5917ab75b83fc026d187f1b5fc48b4a1ce | 87% | Unknown | 1 |
| 12 | Inkling-small mcp-atlas:Inkling-smallconfig_f75f6d17e3c366b14091c9625505c7c70b115a94c65fc825da57566170ece92b | 81% | Unknown | 1 |
| 13 | claude-opus-4-7 mcp-atlas:claude-opus-4-7 | maxconfig_05a06b37c0a4784f1186582e4860942edb8279ac8f2c9a12392cfa5971436ecf | 81% | Unknown | 1 |
| 14 | gemini-3.1-pro-preview mcp-atlas:gemini-3.1-pro-preview | highconfig_d16487ba4140f90ab8e629b2bd15a10683a8a0b38a5b422addc585163fa8fc2b | 79% | Unknown | 1 |
| 15 | glm-5p2 mcp-atlas:glm-5p2config_9699b17040d7a7e000ce022b080619085249b28ef899665d8db4775a89f737fc | 78% | Unknown | 1 |
| 16 | claude-opus-4-6 mcp-atlas:claude-opus-4-6 | maxconfig_07f910f0cb8559b0faf9b23dcb787e5fdbc4c3f879f396903e7bd95686c37c80 | 76% | Unknown | 1 |
| 17 | Inkling mcp-atlas:Inkling | xhighconfig_40527c8c703ea129249d2e9b7e67687c3ef8098f305b5caec0730ef571e53d5d | 75% | Unknown | 1 |
| 18 | glm-5p1 mcp-atlas:glm-5p1config_b052e66e22d1ee4d2bad78bd6113209e193d9488c4c8167aa1d24acfb8e5813c | 74% | Unknown | 1 |
| 19 | gpt-5.5 mcp-atlas:gpt-5.5 | xhighconfig_f5077b4a3b6a10c5fb4ceb2f793cfdc2cac38a5eb5a489a8a3c9aac250172ca6 | 73% | Unknown | 1 |
| 20 | gpt-5.4 mcp-atlas:gpt-5.4 | xhighconfig_0f3ec8178be2beb726efa9e54d9bbe8b0ebe163b046451b1c0e618665453b129 | 63% | Unknown | 1 |
| 21 | gemini-3-pro-preview mcp-atlas:gemini-3-pro-previewconfig_83f12c7c544421b5ca5d1de95594c62fae3c76cf7eccba162e4debc2541c8e64 | 63% | Unknown | 1 |
| 22 | claude-opus-4-5 mcp-atlas:claude-opus-4-5 | highconfig_3554bcad8ceadb2c86d5eef891263f8fd1aa56bae525786b724d3f03c2d2c253 | 62% | Unknown | 1 |
| 23 | claude-sonnet-4-6 mcp-atlas:claude-sonnet-4-6config_4b53a75c0bf72e6e78ceb7f3710a73def172cecfd1481e27fcd95a9c19efbcf9 | 61% | Unknown | 1 |
| 24 | gpt-5.2 mcp-atlas:gpt-5.2 | xhighconfig_880a3ad28d5bd24cc745c1e6bf260f3d3d196263b0d8cdf285f6e5956e605818 | 57% | Unknown | 1 |
| 25 | kimi-k2p5 mcp-atlas:kimi-k2p5config_fba0ee6e849a5c93b689b9d0e3b26034b06b7a2cf9ee2222d6dd7bae9ed30c23 | 51% | Unknown | 1 |
| 26 | Nemotron 3 Ultra mcp-atlas:Nemotron 3 Ultra | thinkingconfig_90a7dd9d9a4e3539f3d56132efa93b8433610735bd9985371cdf6cb9d9c90879 | 48% | Unknown | 1 |
| 27 | gemini-3-flash-preview mcp-atlas:gemini-3-flash-previewconfig_c55ef9457cb3ed93270f0f5fd9bde56665511b7d532751c17944fee26cfff272 | 46% | Unknown | 1 |
| 28 | claude-sonnet-4-5 mcp-atlas:claude-sonnet-4-5 | thinkingconfig_6bff94df52f572fd58c975cffac1e3f415ea59f18493e9d86d141a7a4eb0b839 | 40% | Unknown | 1 |
| 29 | glm-4p7 mcp-atlas:glm-4p7config_4f4efb7aa85ec2fa61d79749a4c2219af212f5fd4df9f0b743c3ac807193597b | 37% | Unknown | 1 |
| 30 | gemini-3.1-flash-lite mcp-atlas:gemini-3.1-flash-lite | highconfig_42937da441ce651eb4640549352dc34c2ff35e1c9d9aa539e3fab35e9a28ad9d | 35% | Unknown | 1 |
| 31 | gpt-5.4-mini mcp-atlas:gpt-5.4-mini | xhighconfig_5131da2f10f8f911dabcc4468860eda25d94ced53ea86759194bac58c9236b14 | 34% | Unknown | 1 |
| 32 | gpt-5.1 mcp-atlas:gpt-5.1 | highconfig_5cc4d6801b3c13b8d9033fac9aa87f665bd0a7383b61f77e8fbb0c1de9892ed9 | 21% | Unknown | 1 |
| 33 | o3-pro mcp-atlas:o3-proconfig_e8fdd869d71f2e1bc97d8ea15b2b753bc9b579a1aad2976d16370aca6735df0c | 9% | Unknown | 1 |
| 34 | claude-haiku-4-5 mcp-atlas:claude-haiku-4-5config_0980c78f9c8c89f5a9484df46ff794b35518b66cbdb0a81dce709b25a7594c9d | 0% | Unknown | 1 |
Explorer evidence
tau3-bench Banking Knowledge
Sources differ; shown separately, never averaged. Observed ; fresh until .
| Rank | Configuration | Capability | Uncertainty | Families |
|---|---|---|---|---|
| 1 | Qwen 3.8 Max qwen3-8-max_sierra_2026-08-04config_a26a8d10f6a1f8c55b2ae39ab85459ab0c65f1c000c357819e67050c18dcc2fc | 55% | Unknown | 1 |
| 2 | Claude Opus 5 claude-opus-5_sierra_2026-08-04config_47e4a643a1c879897e35acc19a52271766e0d24716490d433f47f1250db796f5 | 49% | Unknown | 1 |
| 3 | Grok 4.5 grok-4-5_sierra_2026-08-04config_f4ada0ce1a4b8539b6b16eb70ecd5937576c5612b57153c438fab65dc2644d96 | 48% | Unknown | 1 |
| 4 | GPT-5.6-sol gpt-5-6-sol_sierra_2026-08-04config_a3711cbeeacfc54666ebdd477195ebcaac4642a7aaff878de8a40aa19416a568 | 47% | Unknown | 1 |
| 5 | GPT-5.5 gpt-5-5_sierra_2026-05-05config_ad5041576a055588650eb979cd38244d3c92d872d052d82dafb205bfa74cad3d | 45% | Unknown | 1 |
| 6 | Muse Spark 1.1 muse-spark-1-1_sierra_2026-08-04config_9e9d0ff85e658a60ac616eb6536f77d57ceb67178820283b459b21fc91e37164 | 40% | Unknown | 1 |
| 7 | Claude Opus 4.7 claude-opus-4-7_sierra_2026-05-05config_442ebf6d5800bf504193b385c4554a54f85fd95ef7845c2c54708c68861e0216 | 40% | Unknown | 1 |
| 8 | Claude Opus 4.8 claude-opus-4-8_sierra_2026-08-04config_53dbefe80b9536d86ae8cb24809c569e4c07be1cca283343280e6be8bb563145 | 40% | Unknown | 1 |
| 9 | Claude Fable 5 claude-fable-5_sierra_2026-08-04config_7bf49a074c69d34202fcdb4efe6bd1931566fb4e29b9f6e773de585366ecc67b | 40% | Unknown | 1 |
| 10 | GPT-5.4 gpt-5-4_sierra_2026-03-25config_81277ce76a2443b26ca7f3db93d65526b5c66a673148fbe228d16db598379eb7 | 39% | Unknown | 1 |
| 11 | Kimi K3 kimi-k3_sierra_2026-08-04config_0ac9cb2463f2a18ab6613660534e55deb75861bf471271f6aac39d02e4fbc4f8 | 37% | Unknown | 1 |
| 12 | GLM-5.2 glm-5-2_sierra_2026-08-04config_ebd02a3bcc5de5fdc15841e5214662ff73054096aefd3b9db9de3c1ced3cc679 | 37% | Unknown | 1 |
| 13 | GPT-5.2 gpt-5-2_sierra_2026-02-26config_0a61de28f7e96ed842d0988f4ff17d52f778051137caca3b12c8809fab3a1e24 | 32% | Unknown | 1 |
| 14 | Distyl ButtonAgent distyl-buttonagent_distyl_2026-03-25config_91e6242dafc69cf6d5b045dfa3e30dbca4c38c586b61a1ecbb28962e03b9d7c6 | 31% | Unknown | 1 |
| 15 | Claude Opus 4.6 claude-opus-4-6_sierra_2026-05-05config_79e8f1ddf8bcb22e452c0322f0a78c5fbcba059bfa5acc9acce62e5702eee472 | 27% | Unknown | 1 |
| 16 | Gemini 3 Flash gemini-3-flash_sierra_2026-03-02config_8095fa144e2c9e0944193c86724a27a3b80ddc5fb3d62d561fa298fa54c79701 | 27% | Unknown | 1 |
| 17 | Gemini 3.1 Pro Preview gemini-3-1-pro-preview_sierra_2026-05-05config_03d738b190a028db163266b395b4eca0e0f719f7bb43bd1a6db305d10f91fe2d | 26% | Unknown | 1 |
| 18 | Claude Sonnet 4.5 claude-sonnet-4-5_sierra_2026-02-26config_12ecf4a9f0ae1e2f9815530d97599aef123564d64fddecf8fd39aad2165b1a6e | 25% | Unknown | 1 |
| 19 | Inkling inkling_sierra_2026-08-04config_4a49384540eefc80eaff83393407119b8902549d0f6a45bed272b60bd4ee4937 | 25% | Unknown | 1 |
| 20 | Claude Opus 4.5 claude-opus-4-5_sierra_2026-02-26config_311ca40d213a79344ba2b618fe7f29fcf71ff8cbaaa8cab632499a1bf6dc01b0 | 25% | Unknown | 1 |
| 21 | Grok 4.2 grok-4-2_sierra_2026-05-05config_62cd4f1736a5056e6cf24954b6e0fe45123ac500dcbe4726080840da176bc5a8 | 18% | Unknown | 1 |
| 22 | Gemini 3 Pro gemini-3-pro_sierra_2026-03-02config_66ed191ca632aad42f3b0c528e2912ccef29a49ac134594c4f0209ac45e77ef7 | 18% | Unknown | 1 |
| 23 | Grok 4 fast grok-4-fast_sierra_2026-05-05config_3835bdd87e33611308716fcc185637e9d15136981937b1d92c25d7ae29775927 | 16% | Unknown | 1 |
| 24 | Gemini 2.5 Pro gemini-2-5-pro_sierra_2026-05-05config_5908158689798bbd956792b26fbe7a5a831a7907a7f811a7e974b7231763d00b | 14% | Unknown | 1 |
| 25 | Grok 4.1 fast grok-4-1-fast_sierra_2026-05-05config_320872c79fbebdc15ccb655f4f54fd38971fc9049a5bdb2a2ce5209a2fc755ea | 13% | Unknown | 1 |
| 26 | GPT-5.2 gpt-5-2-none_sierra_2026-02-26config_414bd51eab2ed44504e875d40d7cfe5b9246c80ca5bb91ca01accabf5625ea97 | 13% | Unknown | 1 |
| 27 | GLM-5 glm-5-think_sierra_2026-03-02config_3f1ad20f11e4674c490744ee115c4bc0c88728936353a52de2d451677b63812f | 10% | Unknown | 1 |
| 28 | Qwen3.5-397B-A17B qwen3.5-397b-a17b-think_sierra_2026-03-02config_54cb0b0fed47395e14ff29324863729b3a12e1f93ad171469ba3b588e56dd177 | 10% | Unknown | 1 |
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
- tau3-banking-knowledge benchmark evidence
Explorer evidence
Toolathlon
Sources differ; shown separately, never averaged. Observed ; fresh until .
| Rank | Configuration | Capability | Uncertainty | Families |
|---|---|---|---|---|
| 1 | GLM 5.3 Flash toolathlon-verified:GLM 5.3 Flash | Default | reasoning=maxconfig_a1aadfcb56405a0d0eb191f5135a9d9e2ccb7dee2c238801790214226a6a78ba | 78% | Unknown | 1 |
| 2 | Kimi K3 toolathlon-verified:Kimi K3 | Default | reasoning=maxconfig_3c68f7e1a7f2486c9ff1b6472b4c236cc5336d62caaede33a34a94c8b18131ba | 77% | Unknown | 1 |
| 3 | Claude Opus 4.8 toolathlon-verified:Claude Opus 4.8 | Default | reasoning=maxconfig_40bbb60a17931a21d62e2e9759806e9ee686d1024058b5e441fa991ef3511175 | 76% | Unknown | 1 |
| 4 | Muse Spark 1.2 toolathlon-verified:Muse Spark 1.2 | Default | reasoning=xhighconfig_9feb5dad2c8a17db8a8a081b11fc4f030ecd8b0058bcdf81d49c2fc78dcf1f0a | 76% | Unknown | 1 |
| 5 | Muse Spark 1.1 toolathlon-verified:Muse Spark 1.1 | Default | reasoning=xhighconfig_b3dd3426004e01579b0fa24d5e2d3af6a1055eea5369d7be8cb7a3bbbc76785f | 76% | Unknown | 1 |
| 6 | DeepSeek V4 Pro 0813 toolathlon-verified:DeepSeek V4 Pro 0813 | Default | reasoning=maxconfig_a442fb1aa154afa78f31f9333e14fbcceeb68d81bc8ff09f12cbf0a76563a7b7 | 74% | Unknown | 1 |
| 7 | GPT-5.5 toolathlon-verified:GPT-5.5 | Default | reasoning=xhighconfig_14a00cacba7d65e688a5c760163b8a5c43f157cd0ac3c74a8908ada4bf75af4a | 74% | Unknown | 1 |
| 8 | Claude Sonnet 5 toolathlon-verified:Claude Sonnet 5 | Default | reasoning=maxconfig_1e92cd632b85b5cae4db98f0747a379d6971e1901e817b6c70ba668a1d4f2581 | 72% | Unknown | 1 |
| 9 | DeepSeek V4 Flash 0731 toolathlon-verified:DeepSeek V4 Flash 0731 | Default | reasoning=maxconfig_088be908339ec2c5d18efb1d89f10f2a589d84061a7df91df36136544658b0fe | 71% | Unknown | 1 |
| 10 | Gemini 3.5 Flash toolathlon-verified:Gemini 3.5 Flash | Default | reasoning=highconfig_7424c21364f1c26da7c03d4246703a4574f36bcf742ec051d026f1a2866292a5 | 67% | Unknown | 1 |
| 11 | Gemini 3.1 Pro toolathlon-verified:Gemini 3.1 Pro | Default | reasoning=highconfig_6387719edef57bba35460d38ec2768197f35735c9d27dfbc49a4dacf8b8749c4 | 61% | Unknown | 1 |
| 12 | GLM 5.2 toolathlon-verified:GLM 5.2 | Default | reasoning=maxconfig_32d1f9502ae274db028b789ff6ea67f263fbea181b3f45dd6ab9a33c40321138 | 60% | Unknown | 1 |
| 13 | Kimi K2.7 Code toolathlon-verified:Kimi K2.7 Code | Defaultconfig_013440bcabc69495c6ad1bfdd9fbf082239f7167ccc9dc6f3a721aacb6144341 | 58% | Unknown | 1 |
| 14 | Kimi K2.6 toolathlon-verified:Kimi K2.6 | Defaultconfig_c6aa2be8da38947cda5c3c1ab5ce5ba57e8dd5b1a105e0734ec8b3d785cf6e9a | 58% | Unknown | 1 |
| 15 | Gemini 3.5 Flash-Lite toolathlon-verified:Gemini 3.5 Flash-Lite | Default | reasoning=highconfig_425f38f920acec6ada6502aa651d29309804f062d7ed03c8eed3c54fa62c02d5 | 57% | Unknown | 1 |
| 16 | Hy3 toolathlon-verified:Hy3 | Default | reasoning=highconfig_446911b3349e7d3b5aad2cacc1d0b27500170fc47e6e2a36c433ae9723f807b4 | 57% | Unknown | 1 |
| 17 | DeepSeek V4 Pro toolathlon-verified:DeepSeek V4 Pro | Default | reasoning=maxconfig_7527b100f7599e56734b703f9af93a6879ba5d71dd9e6ee4c136a08140172023 | 56% | Unknown | 1 |
| 18 | Inkling-Small toolathlon-verified:Inkling-Small | Default | reasoning=xhighconfig_2ed2bf4206c9be5f348242ccb5526dc4a6b0195f657a6b33e34f7be422924e5d | 54% | Unknown | 1 |
| 19 | DeepSeek V4 Flash toolathlon-verified:DeepSeek V4 Flash | Default | reasoning=maxconfig_f8a8eb20465f68c1be43c3ad9d4e9c2d41a80a4425f90ea4071fd5ba0355acf8 | 51% | Unknown | 1 |
| 20 | MiMo V2.5 toolathlon-verified:MiMo V2.5 | Defaultconfig_5fbd46f4fa4376922f5b7517a6d938344057a9542a0428d8b0f62cef6a1b2f2d | 49% | Unknown | 1 |
| 21 | MiniMax M2.7 toolathlon-verified:MiniMax M2.7 | Defaultconfig_60f8d3600b9bcb5b8877cbd43ef898f2662079d65b4a7db91dc9319f8507dee6 | 48% | Unknown | 1 |
| 22 | Inkling toolathlon-verified:Inkling | Default | reasoning=xhighconfig_cb28eb8804eaba125c2e9fce39868f34be7b2e47fb2da8261a4f54b3f73d5427 | 46% | Unknown | 1 |
| 23 | Qwen3.5 397B-A17B toolathlon-verified:Qwen3.5 397B-A17B | Defaultconfig_69d341d1f12b8f7988cf1cbd0f72515cb5fab5ddcb250c6ce72a038de5a66e00 | 41% | Unknown | 1 |
| 24 | Nemotron 3 Ultra toolathlon-verified:Nemotron 3 Ultra | Defaultconfig_e3ea78145d4d26cc32da136c704897302f33b6b7cb7d7beba53fcc66f7be9c5e | 34% | Unknown | 1 |
| 25 | Kimi K2.5 toolathlon-verified:Kimi K2.5 | Defaultconfig_8be3ad7082404c320dce37e31ba992fae9c1f78835fc94cef5b7b4e2504672a5 | 33% | Unknown | 1 |
Not enough independent evidence for a top set
- Calibration has not been validated
- Configuration overlap is below the required threshold
- Independent benchmark-family coverage is below the required threshold